Half-way through this post, you'll seriously wonder why you've spent so much time obsessing with Adobe/Google Analytics/Chartbeat or other web analytics tool.
When you are done reading the post, you'll be super mad that your marketing strategy is not more influenced by your competitor's data!
Such is the power of being able to proactively identify which of your competitor's strategies are working well, where their current customers come from, and what specific tactics you should experiment with to create and advantage for yourself.
Regular readers of the blog know of my deep love for competitive intelligence analysis. My first blog post on the topic of CIA was on 14th Aug 2006! Competitive Intelligence Analysis: Why, What & How to Choose. CIA also formed one of five foundational elements in my best-selling book Web Analytics 2.0.
Since then, as luck would have it, we have more tools, they are smarter, and have richer data-sets.
In this post we'll go back to the wondrous world of competitive intelligence analysis. We'll look at an incredible cluster of examples, from the simple to the sublime, that will help you learn practical strategies you can immediately go back and apply in your role as a Marketer, Analyst, or the Boss of all Things Digital.
Here are the key elements we will explore in our quest to become CI Analysis Ninjas:
+ Foundational Concepts/Caveats
Competitive intelligence data will never match your site's analytics tool
Competitive intelligence tools will never match each other!
Small sites are out-of-luck
Small countries are out-of-luck
Site-centric CI vs. Ecosystem-centric CI
Tools used in this post
+ Traffic Trends Key Metrics Analysis
+ Visitor/Audience Type Profile Analysis
+ Traffic Sources Competitive Analysis
Search/Keyword Competitive Performance
Display Advertising Opportunity Analysis
Social Strategy Effectiveness
Geographic World Domination Opportunities
+ Onsite Behavior Key Metrics Analysis
+ Content Consumption Competitive Analysis
When I say this is the ultimate guide for learning how to do amazing competitive intelligence analysis, the above list should clearly demonstrate that I'm not kidding around. You are going to learn A Lot!
My recommendation is to make sure you read the foundational concepts/caveats first (and twice). Without it, you are not going to shed your imprecise assumptions/impressions/past beliefs, nor learn about new possibilities. After that first section, you are welcome to jump around if you would like to, though I do recommend reading one complete section, practicing it, and then moving to the next one. Bookmark the post, keep coming back to it.
Excited? I am so excited, come on!!
Foundational Concepts/Caveats.
It is incredible to realize that you can spy on your competitors. While not quite as sexy as actual spy craft, it is fun and insightful. But, there are some important things that you should be aware of/brush up on/deeply internalize. Or, you'll either abandon this fun journey too quickly or, worse, stay on it and suffer way more than you need to.
The post outlines eight sources vendors use to collect data: Toolbars, panels, ISP data, search engines, self reported, scraped/indexed, hybrid, and external voc. The value of competitive intelligence (CI) is huge, it is well worth investing in learning how the sausage is made. You'll suddenly understand why, for example, something like Alexa is a unwise investment of your time.
Always ask your vendor two questions: 1. How do you collect data? 2. What is the sample size?
Multiple sources, of the eight above, are better. The larger the sample the better.
Competitive intelligence data will never match your site's analytics tool.
Here's how I've often explained the two different data-sets. In one case, Charlie's standing at the door of your supermarket and every time a customer walks in, Charlie presses a button and records that entry. In the other case, Pooja is standing on the moon with a telescope and each time she detects that someone's walking into your supermarket, Pooja will press a button and record that entry.
Since Pooja's likely to miss some folks walking in (gosh darn clouds!), she will look at all her observations of other supermarkets on Earth, on other planets, and do her best to give your the best normalized, indexed, algorithm'ed number.
You, see why the two numbers might not tie?
If you do not possess inside you the capacity to see beyond this, it is important that you stop reading this post and head over to buzzfeed because that might be a lesser waste of time (though that is giving buzzfeed a lot of credit :)).
CI data is valuable because…You'll look at the same competitive space you live in. The numbers for you and your competitors will be roughly equally wrong. You won't look at specific differences (it says 3,256,293 for me and 3,256,193 for our competitor, we win!), you'll compare long-term trends. You won't obsess about the long-long tail (data that sparse is not confidence-worthy), you'll look for big fat head opportunities. And so on, and so forth. In all of those cases (outlined below), you'll derive amazing, actionable insights.
Perhaps thinking about CI analysis in this way will help… You are comparing slightly-bruised-apples to slightly-bruised-apples. And that is 100% ok. But if you can't get over Charlie not matching Pooja, I've mentioned the buzzfeed option above.
Competitive intelligence tools will never match each other!
A slight variation of the above.
Each tool uses their own data collection methodologies, their own data processing algorithms, their own assumptions, their own way to plug in any holes, and their own ways of "normalizing" the data-set to give you the best possible answer. They will never tie to each other (if they do, run, the zombie apocalypse is upon us!).
Here's a simple example.
This is Occam's Razor's data using a tool I really like and use a lot, SimilarWeb…
It is a profound over-estimation of reality.
Here is the data for this blog from another tool I really like and use a lot, Compete…
It is pretty close to reality, though everything from Dec '13 through mid-year is imprecise.
And here is another tool I'm sure you've heard of Hitwise…
It profoundly under-estimates my traffic.
This does not mean these tools are useless.
Hitwise, for example, has one of the largest datasets of this type in the world. It is the differences between how they collect and process the data that creates the differences.
My recommendation is to investigate many tools, and then stick to the one you are most comfortable with. You'll be surprised, especially by the end of this post, how much better your digital strategy can be from comparing slightly-bruised-apples to slightly-bruised-apples!
There is one other thing going on in the three pictures above: All CI tools are pretty terrible at data they report for small sites (and heartbreakingly Occam's Razor is a small site).
It is a matter of how the data is collected. Their listening posts simply can't accommodate for the low signal in all the noise, no matter how many of the eight CI data collection strategies they use.
Here's a rough guide from my experience (as you do this more, you'll form your own filters).
If you have under 100,000 Unique Visitors each month, your data (or your competitor's) in any CI tool is mostly useless.
If you have between 250k-500k Unique Visitors, you are in a decent place.
Any site with more than 500k, to say 1 mil, is in a very good place.
Anything over 1.5 – 2 mil, you should start to see patterns in the CI tool you see in your own analytics data.
So, what do you do if you have less than 500k Unique Visitors each month? Do analysis on your big competitors! Learn from them (specific recommendations below). And, for now rejoice that your big competitor has nothing on you! :)
Small countries are out-of-luck.
Similar reasons. If you are in Croatia, you might not have very good CI data for your country (you'll have to make do with relaxing on your world-class beaches that are the envy of the world!).
Sadly, the same for perhaps the entire African continent. Central America. Others.
Sometimes there is no signal in all the noise and if you do hear a signal, spend time asking the two questions in the data collection section above.
So, what do you do if you are in a small country? Learn from what is happening in a like-type country larger then yours.
If you are too small, use industry sources to do ecosystem-centric analysis (more on in the last section of this post). If you are too small, maybe use surveys or primary research to learn competitive examples.
Site-centric CI vs. Ecosystem-centric CI.
As Edward Snowden and the NSA/CIA/MI6/GCHQ documents releases have shown, there are many kinds of competitive intelligence in the world. Our interest here is not that type of intelligence. [Note to NSA reviewer of this blog post: I'm a harmless digital marketer, totally not worth your attention.]
In our scenario there are two types of CI analysis.
Site-centric analysis is primarily focused on looking at direct competitors, or adjacent ones, and learning from what is happening with Acquisition, Behavior and Outcomes on their website (desktop and/or mobile).
Ecosystem-centric analysis is primarily focused on what is happening in the industry, in the sub-category, the small or big niche that everyone's playing in.
Both are valuable. In this post we'll focus mostly on site-centric competitive intelligence. I'll close with a fast sprint through ecosystem-centric analysis awesomeness, but it is not the main focus of this post.
In this post I've used Compete, Hitwise, SimilarWeb for site-centric analysis. Google Trends, AdWords Keyword and Display Tools, Raven Tools, for ecosystem-centric analysis.
The common thread across all these tools is that I like them, regularly use them, and, if they are not free, have access to the Pro version.
The two exceptions are… I tend to dip in and out of my Raven subscription (primarily because I tend to dip in and out of SEO analysis needs). I tend to use Hitwise a lot less, or not at all, because it is a very hard to use, it is slow, the UX/UI, metrics and reports have not really evolved over the years. This is heartbreaking because the data-set, as mentioned above, is nice.
The other reason to use the above tools in this post is that you can get some stuff for free (Compete, SimilarWeb), or the full version at an affordable price, or you can get full version trial for a short-period (Raven), or you can get all the data completely free (Google Trends, Keyword and Display Tools).
There are many, may other tools in the world, wonderful tools. Please try them. My goal here is to only talk about what I know, use and like.
Important note: Compete is the only tool in this list that shows US Traffic only. It does not show data for non-US visitors to your site. And, if you type in an international site, like http://www.dhl.de, the data you are seeing is only for US visitors to dhl.de.
That is a lot of foundational elements! I'm sorry, but it is a complex topic and I want to make sure that when your boss questions you, or you are quizzed in a job interview, you are extremely well prepared to rock on this subject.
Enough talking, let's go do some analysis!
Traffic Trends Key Metrics Analysis.
The very first thing you are chomping at the bits to know is how much traffic your competitor gets! Easy question to answer, and free with Compete (remember, this is US visits only) or SimilarWeb (global traffic). So let's start there.
I've chosen to focus on one company for the entire blog post, for the sake of consistency and to make the lessons stick a bit more. We are going to look at Getty Images. I know some folks there, super-lovely people. Thought they would like to see an analysis of their digital strategy, and that of their direct competitors.
The first thing you'll do any CI tool is you'll likely look at your own data. I know you can't resist the temptation, go ahead, look at your data… here's Visits and Average Stay (Time on Site) for Getty in Compete…
Ohh… nice lines! You can look at the trend, see how it's going up and down. I also know that you can't resist comparing it your GA/Adobe data. I did recommend not doing that, but go ahead. I'll wait here. Happy now? Probably not. Are you going to listen to me now? Let's move on to doing productive things.
Remember, the best use of CI data is to contextually compare. Pretty much all tools have a button called Compare, and when you press that button you'll get something similar to this…
I really like this feature in SimilarWeb and Compete. They'll both tell you who they think the competitors are (sometimes you see surprising one, and that is awesome!), and how closely they believe the businesses overlap.
All I do is add the top ones from the list, press Compare and…. boom!
So much more fun. (You'll see my bias above, I prefer to use Unique Visitor comparisons rather than Visits.)
We (Getty Images) get very valuable context for our performance. We looked pretty good by ourselves in the first chart, now we understand how we are doing compared to a direct competitor-set.
Getty is the fourth green line. Fotosearch is tiny, and ignorable in this context. So we are dead last. Many times lower than Shutterstock. Good thing Getty also owns istockphoto (orange line)!
You can see that everyone's doing better recently. Right under the graph above, you can see some helpful raw numbers for recent performance. This can possibly make a good module in your tactical dashboard (we'll see cooler stuff we can add to our strategic dashboards later in this post).
Allow me to share an important tip. Every website in the world has a unique business strategy. Even direct competitors like Expedia and, say, Orbitz. It is important that you take this into account. Is Getty really competing with these sites, given its business model? Is it someone else? Or is this a gap in their business approach they should fix in some other way? These are good questions to ask when we undertake these comparisons.
So, what do you do with this data?
1. Compare long-term performance between competitors? Identify surprising peaks and dips, look for reasons why. (More on that below.)
2. Use this data to create benchmarks you can use for providing important context to your CEO when she/he looks at your own performance. Identify changes in your marketing mix/strategy.
3. Look for surprising competitors suggested by these tools. Why are you not worried about them?
Quite straightforward. Good way to dip your toes into the water.
Visitor/Audience Type Profile Analysis
Once you know how traffic trends are playing out, and you want to crush your competitors, you are surely curious about what kind of people are visiting your website vs. theirs.
No problem. You can look at the demographic, psychographic profiles.
Most competitive intelligence tools will have data that will allow you to compare Age, Income, Gender and other dimensions easily. This one's from Compete and I choose us and Dreamstime (as they are doing 2x better than us!)…
You can see that there are clear differences between the two sites. Remember my call to action on ensuring you understand your business strategy, this is where you can see it manifested the most.
Other tools, can provide more dimensions including children, political leanings, ethnicity etc…
You can see how this might be a lot more interesting for a magazine or content sites, and perhaps a lot less to other types of businesses.
On that note… If you search through the 300 posts I've written thus far (this one is 301!), and the 638,442 words in those posts, you'll notice that I've never, ever, never asked you to use demographic, psychographic data to do anything with digital. This is primarily because I fundamentally believe that this is almost useless data. On the web we can infer a person's intent using their actual behavior. Why do we need to know how old they are or if they are republican or female? If I can discern intent why do I need demo/psycho data? If I can infer that you want a Nexus 9 or are looking for an article on Libya or want to learn what a Roth IRA is, why do I care how old you are or if want to school? Using your intent, I can deliver the answer you are looking for! But, vice-a-versa is not true. If you filter by demographic/psychographic data, you will never engage a 75 yr old woman looking for an iPod (because you would already have ruled her out as old!). You can get inferred demographic data even in Google Analytics. It is substantially less useful than you might believe.
In the old world of TV/radio and no behavior data you had little choice. You had to rely on 300 surveys filled out by badgering people on street corners or when they are having dinners at home. You made do with what you had, and that was ok. But now, you don't have to do that anymore. It is 2015.
I'll get off my high-horse now. It is ok if you don't agree with me. I just want you to know why I believe what I believe.
What I find a lot more interesting (behavior!), is data like this one from SimilarWeb….
It helps you understand the web wanderings of people who also visit your site (or a competitor's). What categories of content do they consume? What topics are covered in that category? What other websites do they visit?
I find this behavior to be a million times more useful than demographics or political leanings. And I'm totally ignoring that demo/psycho data is deeply inferred and algorithmized.
It is a lot less useful to know that 22 yr old, gay, African-American, republicans, who did not go to college, and have four children are 7% of the audience. It is a lot more useful to understand, this is what they search for, this is what they read, these are the sites they tend to visit.
So, what do you do with this data?
1. Your boss will absolutely love the demographics and psychographics the first time you show it to them. Do it. Earn some love. Why not? Then, slowly, try to wean them off this non-very-useful data.
2. Look for Audience Interests type data from your CI tool. Create hypotheses as to what this means for your content strategies on your own and rent digital platforms. Structure tests to validate these hypotheses.
3. Identify the impact on your display advertising campaigns. This will have direct implications on your ad targeting, ad content and ad landing pages strategies. Recommend changes.
Done right, you'll be surprised how effective #2 and #3 are. Both at improving your digital engagement and acquisition effectiveness, and at weaning your leadership away from the shiny hollow demo/psycho obsession.
Traffic Sources Competitive Analysis
You know how many people visit your digital existence, you have a better idea about who they are based on their personal attributes and their behavior.
Time to understand where they are coming from!
This is the biggest section of this blog post because it is often the area CI vendors will have most direct knowledge about people's behavior. Remember our metaphor above? Pooja is standing on the moon with a telescope and observing your supermarket, she will have some knowledge about what's going on inside the supermarket but she will know a lot more about where people were before they came to your supermarket, where they went after (she can see the trails more clearly), what search keywords brought them there, etc.
There is a lot of gold to mine here. Hence, a lot of your time will be spent in acquisition analysis to improve your digital advertising and marketing strategy.
The highest level acquisition view you can get is knowing where people come to your competitor's site from, and where do they go after their visit. (For your own site, you don't have Outgoing data in your web analytics tool, so it is a nice to have from a CI tool.)
Look for competitors in the Incoming list (this also, won't be in Google Analytics or Adobe or WebTrends). Look at everyone and everything in the Outgoing list. If people are really happy with what they get on our Getty Images site, why do they go to istockphoto and Shutterstock next?
A great use of this data is to compare and contrast (with CI tools, because of all the caveats at the start of this post, it is good idea to always compare and contrast!). For example, what are the differences between Incoming sources for Dreamstime and Shutterstock?
You can see clear differences between the two companies. They might have different business models, but this should be great external third-party data to take to your CMO and have her/him ask questions like: "Why is it that our second biggest source is Bing when our competitor's is Facebook?" There might be a good reason, but it is worth asking the question. Or for Shutterstock to wonder why Dreamstime seems to be so, so, much better at Google! (More on how to discover why that is the case, below.)
Naturally you want to dig below the surface. My favourite method is to use the Traffic Dashboard report in Compete (as I'd mentioned in my post on data visualization strategies and examples).
This is what it looks like, for our website and our images category leader…
You can notice big differences in acquisition strategy right away – from the size of the boxes, and from, for example, seeing that Web Portals and Search is darker green (more growth) for Shutterstock than for us.
Beyond that, clever analysts that you all are, you'll notice subtle things like Shutterstock is likely better at building loyalty and getting people to come back again and again directly when compared to Getty. And many other such things are waiting to pounce on you!
If you are just starting out, you probably just want to know, what's their Miscellaneous and how can I steal more of that! Good question.
Just click on that box in the Treemap, and you'll get the answer…
Sweet, right? You can see the top sites, and you can see if in the last month traffic from that source grew or fell.
For me, the misc view is always a great place to look for relevant sites to build relationships with for acquisition (links, display ads, content, something else).
Another favourite of mine is the box below the Treemap in Compete where we can see cute things like these…
Saves me the effort of actually doing my job. :) It helps me find the top growing, declining, long-tail (#omg), social and search visits.
So, what do you do with this data?
1. Make big picture recommendations about changes to your company's owned, earned and paid marketing strategies. Stop wasting time with Social (or do more!). We need to re-evaluate our YouTube strategy. We need to shift our strategy to more dramatic focus on SEO. Etc. In each case, frame it as a hypothesis, test it, make bigger changes.
2. Identify tactical recommendations for your display, relationship-building, and other such strategies. Be specific, back it up with drill-down data from the Treemap view (or equivalent), and top movers (up or down).
3. This is much harder but try and figure out the implications of the Outgoing sources list. Are people going to that list of sites because they did not get what they want from us, are we more expensive, is our internal search broken, do we have the world's worst checkout experience (which is true of one of the sites we are talking about today)? This does require a much more strategic approach, master #1 and #2 before you go here.
Most advertising and marketing strategies are based on what we see in Google Analytics, or what we feel might work. Both have value. Additionally use CI data. Use what's already working for your competitors.
Search/Keyword Competitive Performance
Once you are done with the big traffic sources, it is time to focus on the source that is usually the single biggest: Search!
We will follow the same pattern, start macro, dig a bit, dig a bit more.
The very first thing you would want to track are the overall traffic trends. In this case, a simple distribution of overall search traffic and the amount that is coming via paid search. We are using Compete for below report, you can get a similar one in SimilarWeb (see what I did there? :)).
If you've learned one lesson thus far in this post, your site's data by itself is not all that interesting. You want context.
So here's us and the leader in our space…
Interesting! We were a bit closer last year, we have both grown since then, but our competitor has shown much better growth since then. Boo!
The next question is obvious, let's take a look at the traffic from paid search vs. organic…
Big difference here. You can see a specific shift in strategy starting May 2014 for Shutterstock, while we have continued to plod along at the same pace. Ouch.
I believe that it is not all that important if your competitor is beating you with paid search or organic search. If they are beating you with search, you have to take action! I've heard many a senior business leader trot out the excuse but they are beating us with paid, we don't have money to spend! My reply to them is: If search traffic is profitable for you, and PPC delivers profitable traffic, why would you let your competitor win? It is really that simple, and yet so hard for some executives.
If you compare the two graphs, Shutterstock is beating us in search overall by approx 1.5 million Unique Visitor. The delta in the above paid search graph, around 300k, is just a part of their winning strategy (which clearly is SEO + PPC, as it should be). We have our work cut out for us.
Then, defying every excuse from our executives for not doing the right thing, we shift to competitor crushing mode!
The first stop should not be the tsunami of keyword data every tool will provide you. In your CI tool, try to look for a feature that will allow you to filter and focus.
In Compete this is using the various filters they've already created. High traffic keywords. Keywords that bring a lot of people who also spend a lot of time on the site. Keywords with high organic share. So on and so forth…
It helps bring a lot of initial focus to your efforts. And trust me, you will need it.
Once you identify your criteria, you can dive deeper into the data. The columns might look a little different, but in general this is the type of information you'll see…
It is easy to figure out what the key phrases are, why the competitor is winning, and how much more valuable one keyword is compared to others. I find both the paid and organic (or natural as Compete calls it), and the Average Time Index to be valuable in drawing up my initial world domination plans.
Similar report from HitWise. It does not have time, but you can see the other metrics are roughly similar…
I recommend taking the above mentioned keyword research approach first, learning from your couple biggest competitors, and doing this one second: Death match!
You type in your site, you type in your competitor and before you can say Tom Brady, you've got data!
I really like that it shows if you win on volume (amount of traffic you can get vs. your competitor) AND gives you a sense for quality by sharing how much time that traffic spends on the site.
So, you can see that there are some cases where we, Getty, are winning on volume, but for this bunch, we lose every time when it comes to getting people to stay on our site (even when we win!).
It is quite possible that Shutterstock is the suckiest website in the world and the time advantage exists for them because the site is so much more painful to use. It is easy to check, type the keyword into Bing, go to their site, see the experience. Now that you know how good it is, feel shame for coming up with the excuse for losing the time advantage, and resolve to win by making V for Vendetta (!) your new mantra. :)
Now you know both keywords to include into your SEO and PPC awesomization strategy, and the landing pages and experience to fix. How lovely.
In my quest to make our lives easier, I wanted to share a couple other features I adore.
In SimilarWeb, I love this little button you'll see on top of the search terms report: Exclude Brand Keywords.
It is so useful! The biggest opportunity often is not in the brand keywords, it is in the category and long-tail keywords. And these reports are so often clustered at the top with 152 variations of the brand names. I hate it so much because I have to manually filter to find something useful.
Now, I just click on that button and magically, with no more clicks, most of the brand terms disappear. You are left with the category keywords I'm so interested in, and some mis-spellings of the brand terms which are much easier to deal with.
A second feature that is truly awesome, is in the Competitors folder. It's the Keyword Competitors report.
The first tab shows the businesses you compete with in organic ranking, people taking away your God given right to capture 100% of the clicks for all your brand, category and long-tail terms. :) Here's the report for Getty Images…
It is so nice to have this data so easily accessible.
You'll see some obvious things, but you'll also notice clever things. Shutterstock is #7 in organic search competitors list for Getty Images. But Getty is #10 for Shutterstock! The top ten competitors for Shutterstock only have three in common with Getty. Hmm… weird. These types of analyses help you understand who your audience is, and that of your competitor. It also helps you figure out how to make very specific competitive choices about who (/what cluster) you will focus your SEO strategy on.
We often make those choices blind (or worse, based on what the HiPPO believes). At best it leads to less winning, at worse it causes your business to suffer big time.
The second tab of this report shows you the same data for paid search…
The first one is Gettyimagesbank.com, the Korea based part of Getty. This should be an easy check, if they are really competing with each other, or apply segmentation to find geographic root-causes.
I'm sure you'll notice that the list in paid competitors looks different from the organic competitors. This is extremely common. Yet, most companies rarely have a search strategy that accommodates for this reality. Now, you don't have to be one of those companies.
You can also identify valuable insights like you compete a lot more with some companies in paid search (for example, Fotolia and Offset), and you compete with different ones in organic search. Now, you can have a smarter strategy. You can click on each of the sites and drill down into their detailed data, and from there to their own deep search analysis. And discover things like Offset is owned by Shutterstock! #arrrrhhh!
So, what do you do with this data?
1. Create benchmarks for Overall and Paid search traffic patterns of your competitors to create new and improved goals – month-by-month – for your own macro search strategy. Use these to make specific recommendations for search budget (to include people, process changes and ad spend).
2. Change your head and long-tail SEO strategy (links, content, distribution) based on your analysis of not just Share of Search or Volume metrics for keywords, but using engagement/behavior metrics to ensure your target list is pure and refined gold. Deeply analyze your organic competitors to get smarter insights (rather than random we think we like these keywords or you should rank for those other ones based on content you have – forget content you have, focus on what's working for the competition, let that have a say!).
3. Identify a very precise list of paid search keywords using both your paid competitors, and their most successful keywords. Ensure you include opportunities identified by your current head-to-head death match with your top competitors. Remember, winning paid search should include not just keyword lists but also a reflection on what ad targeting is working, the ad content and landing page strategies.
Search continues to be one of my absolute favourite ways to win short and long-term. (Email marketing is another fav.) There is perhaps no richer competitive intelligence dataset than for Search. Use it. Especially important in a world of Not Provided where your web analytics tools are empty.
[Bonus: Keyword Analysis Options In The Not Provided World]
Display Advertising Opportunity Analysis
Most competitive intelligence tools don't have anything for display advertising. This is sad, primarily because if you create great display ads you have a near amazing ability to engage potential customers at every stage of the See-Think-Do-Care universe. Not only that, with all the power that is coming to fore with programmatic buying, you can get so very close to getting the right message to the right person at the right time.
Only people on their way off the cliff now execute digital display strategies of the past which can be summarized as: Spray and pray!
With SimilarWeb you can get a sense for how your competitor is doing with display advertising. You can get the overall traffic numbers…
Which gives you a sense for how strategically you are treating this channel, and how your competitors are doing it.
In isolation our graph looks like are on a rocketship! But in comparison to Shutterstock, even after they seemed to have pulled their pedal off the accelerator, you can see we are at least 40 times smaller!
From there you can dig a bit deeper to understand what platforms your competitor is using, and what categories they are targeting their display ads…
And from there you can drill down into each category to understand what domains they focused on.
As usual it is less that you have to completely copy their strategy, and more that this gives you a good idea of what they are doing, and often what is working for them. You use that as context to create your own strategy.
So, what do you do with this data?
1. Quantify the implications of the gaps in your current display strategy, or lack thereof. The first graph should do fine.
2. Create a specific and targeted list of categories and domains you should target influenced by where your competitors are (and where they might not be). Especially analyze the companies beating you at Paid Search. Also leverage the list from the Compete Treemap view of incoming sources. For each category, you need to partner with your creative teams on ad content, with your marketing team for ad targeting. That is how you avoid spray and pray.
3. If you want to get your company to use display advertising in a smarter manner, and not fall for old cheap tricks like View-Thrus, it is critical that you understand Multi-Channel Attribution Modeling and ensure that you are using both Assisted Conversions and Last-Click Conversions to measure success. Oh, and if you are using both Macro and Micro-Outcomes as well, be careful for I just might run you over with a big hug and a thousand kisses!
Pretty straight forward. Even if it is a lot of work, across multiple teams.
We can't forget Social can we? That's what the hype is all about. Even if organic social exposure on top channels are approaching sub 5% of your followers!
I'll recommend this report from SimilarWeb to be the cornerstone of your strategy…
It highlights a head-to-head comparison of how you are doing vs. your competitor at a macro level and on specific social channels. The metric being displayed is Visits.
While I like the top right (you can get it from the traffic sources section above as well), I really adore the bottom part of the report. It shows Traffic Share and distribution of the share.
If you click on one of those referring social pages, you get the actual pages!
This is quite helpful because now you can click on these links and understand what they are doing on YouTube specifically that is driving traffic to their site. Or on reddit or vk.com. Awesome sauce.
So, what do you do with this data?
1. Identify channels that are more effective for your competitors. Contrast this with your strengths. Make re-allocation of people/money recommendations to your leadership team.
2. Content analysis via use of the second report above is perhaps your bff in trying to change your leadership's mind to shift from shouting to conversation marketing on social channels. Deliver five specific social efforts that your competitors have rocked, that you could possibly emulate (or do one better).
3. If your data looks like above, you'll quickly realize that you are among the 99% that stink at using one of the best social channels on the planet for businesses: YouTube. Leverage my Reach, Build, Engage framework for YouTube and identify current weakness in your strategy and recommend fixes.
Social media continues to be ROI positive on the long run if you are willing to invest in conversational marketing. If you are looking for short-term revenue or quick wins, you are better off skipping it (do something boring, but profitable, like Email or Affiliate marketing).
Geographic World Domination Opportunities
We've looked at so many dimensions related to where people come from. One last one remains, geography.
Most tools have this data available, here's how it looks like in SimilarWeb. I've taken the liberty of pasting two reports together…
There is not a lot more interpretation required. Where are you winning, where is your competitor winning.
Between direct competitors the level of stark difference you see above is rare. There seems to be little on common between Getty and Shutterstock. Even more surprisingly, the US traffic seems to be so low for Shutterstock. For Getty, this should lead to a big review of the current marketing strategy. Especially since Shutterstock is 4x the traffic size as Getty.
The world we live in is interconnected. But, more than that, the cost of providing digital only services (like photos, or literally anything in the cloud) is extremely low. There are few remaining decent reasons for any company to be country of origin centric.
So, what do you do with this data?
1. Provide a list of target countries to your management team where new opportunities exist (as identified by your competitors). At the minimum, make recommendations for decent sized experiments to validate profitability.
2. Do a little competitive intel of your own. For just a few dollars a month get VPN software. Switch your IP to the first opportunity from #1. Visit your competitor's site. Look at popular categories. What is selling more/less (usually available in a straightforward manner). Identify specific categories your company can enter into initially. (You can do this for any type of website, not just photos.)
3. If your CI tool provides City/DMA level information, use that to be more precise about where to target your advertising in these new opportunity geographies.
This analysis is not very complicated. Usually requires some grunt work.
Notice in the screenshot above, you not only get traffic data, you also get other delightful behavioral metrics. Good time for leaving acquisition behind, and moving on to onsite behavior.
Onsite Behavior Key Metrics Analysis
No matter how powerful the CI tool, observing deep behavior inside a website is hard. Unless the site has tons and tons of traffic, the these metrics can be a guess. See my guidance about traffic size and confidence at the start of this post.
Usually you will get four or five behavioral, what happens after a person lands on the site, metrics in a CI tool. Bounce Rate, Time on Site, Pageviews per Visit and Visits per Person. Here's a composite of Compete and SimilarWeb so that you can see how it will typically look…
Visits per Person is particularly difficult to get. It is nice to have that as a measure of loyalty to the site.
(I'm pasting two different tools above to remind you, again, that this data rarely ties. You need to stick to one and then…)
Always make sure you contrast the key metrics with that of a competitor! Above data is for Getty, and these numbers for for Shutterstock…
You can not the raw differences, and also, on the right, how these trends are performing over time (month/year/longer). That is where the key insights are.
Here's a screenshot of behavior metrics compared across multiple competitors in SimilarWeb…
What is up with Fotosearch's bounce rate! Yes, yes, their bigger problem is traffic, but still.
So, what do you do with this data?
1. Look for big deltas – positive or otherwise – in behavior performance. The numbers won't have too many insights for you, hence visit your competitor's websites and identify reasons for why Visit Duration or Pages/Visit might be long or short. Use their top search/display/social campaigns to visit their landing pages to identify why bounce rate might be high or low. Use these to make recommendations to your user experience team.
2. Time on Site and Pages/Visit are a bit dicey, you never know if a high number is good or bad. But, you can use Bounce Rate and Visits per Person to create benchmarks to index your own performance.
3. It is particularly useful to look at behavior metrics a couple months after your key competitors have made big changes to the UI or UX of their site. They probably won't tell you how it's going, but you can come check it out in your CI tool. :) Your insights will be of value to your user experience team.
I'm sure you are wondering why you don't see something like Conversion Rate in these reports. The reason is quite simple: It is hard/impossible.
As sessions switch to https (secure experience), it is difficult for prying eyes to peek through and know that a conversion happen. Honestly, it is hard to even know if someone added to cart. It is not totally impossible. One of the methods to capture CI data is to use intrusive monitoring software (Nielsen, ComScore, others use this strategy). Under this method with some work you can see a conversion. Just be careful about it. Use your good taste and common sense to identify the impact of sample size and sample type bias in such data.
Content Consumption Competitive Analysis
I'm sure the question on your mind is, what are all these Pages per Visit? What is the Time on Site being spent on?
This is where we hit into some limits. It is hard to get page level reports, for the reasons I've mentioned in the above section and at the top of this post.
But there is one report you can perhaps get some value from. The Sub-domains report.
Not every website is organized optimally for us to benefit from this report, but if it is, then you'll get something like this for Dell.com.
It gives you some sense for what content is being consumed, and the amount of change. It might take you visiting your competitor's website to get more context behind these rows of data.
SimilarWeb also provides this analysis. They provide % Traffic Share of each sub-domain – roughly equivalent to Unique Visitors above.
You can also look at the top sub-directories, and drill-down into the top pages in that sub-directory…
The data here, in pretty much all tools, is a bit iffy. But it does give you some sense for popular pages (insperion laptops are more popular than xps laptops).
So, what do you do with this data?
1. Most of the time you'll see that this data is not very useful. But if your competitor has been sweet and organized their sub-directories optimally, it is a worthy investment of time to infer their strategies when contrasted with yours. You might recommend prioritizing certain content higher to your leadership team (both for creation and advertising).
2. If you see anything decent at a page-level, that could also help provide specific insights to your website team.
I do wish we had a lot more here. But, we don't and while you might not agree with me, it is kind of nice that some data our competitors can't get to. : )
#OMG Mobile, Where's Mobile Data!
This past thanksgiving 52% of all online traffic was mobile per IBM. Mobile phones are being used early in the consideration process, they are driving influence. Tablets are driving purchases. Christmas was even better for Mobile (traffic, sales, store visits, everything). The next billion people who come online will only have mobile devices.
I believe that this year mobile will account for a majority of the site traffic (greater than 51%) for commercial websites, and stay majority forever.
I share that as context for the fact that many competitive intelligence tools don't have any mobile data as of now. Or in the rare cases that they do, it is such a small sample size that it is completely untrustworthy. So be aware that this casts a pall of gloom over our CI data.
[This gloom, mercifully, does not extend to the Google ecosystem-centric competitive intelligence tools below. You can segment that data by desktop or mobile. #hurray]
As you use CI tools, and perform the wonderful analysis above, check your site's digital analytics tool for your mobile-desktop split. That should serve as guidance regarding how much might be missing from your CI tool.
Ask every single CI vendor you use: What is the sample size of your mobile data, what is it for desktop? Encourage an honest answer. If you have the power, put immense pressure on them to fix this issue on a priority basis. I'm doing this every other Tuesday.
SimilarWeb does clearly label their data, big props for them for this transparency…
And, this is very better, for each website they have a completely different report for mobile traffic. It provides us with some key insights.
First, you can see the long term trend in mobile and desktop traffic….
Second, they also provide mobile and desktop data split out for the four key metrics that they have in rest of the tool….
This is extremely helpful as you can see huge differences in behavior. You might have expected this difference, but it is nice to see it quantified.
Finally, what is a good CI tool without the ability to compare and contrast. You can create a report to analyze the trend in mobile traffic (Average Monthly Visits) for yourself and your competitors…
You can see that the difference between Shutterstock and all other competitors is the starkest for mobile. Shutterstock is doing so much better. Good food for thought for everyone else. Or a swift kick in the… you get it.
At the moment this is all we have in SimilarWeb (let me note, it is much more than anyone else for now!). As they evolve, hopefully we can see other valuable acquisition metrics and dimensions to enable us to do smarter mobile analytics.
Ecosystem-Centric Competitive Intelligence Analysis
A quick refresh from the very top of this post: Site-centric CI is primarily focused on looking at direct competitors or adjacent ones. Everything we have done so far (above). Ecosystem-centric CI is primarily focused on what is happening in the industry, sub-category, that everyone's playing in.
I'd promised we would take a quick whirl through it (mostly because I would not be happy if you did not know this). Brace yourself for a fast ride!
Ecosystem analysis will start, usually, by looking at the category we are playing in. It is likely you are tired of the Photos space, let's look at something more fun, Jewelry!
You'll initially want to know how the category is doing. Go to Profiles, then Category, pick yours. (One of the 252 Industry Categories is… "Competitive Web Analytics"! It lists 125 websites, but not this one! Boo! Compete! Boo! :))
This is what the report will look like…
You know the broad trends in your site traffic, now you can see how the industry overall does. (Is it surprising that there is such huge peak in Dec but barely anything on Valentine's Day?)
You can also note of the "available traffic" metrics, Unique Visitors and Visits. And are your numbers showing the growth you see there? They should, right?
You can scan the consumption metrics (pageviews and total time spent – 246 million!), but then ignore them as they are quite useless aggregated at this level.
Next, we drill down to understand what are the top x sites in the industry….
You have a much broader view of the world, because your starting point was different. Ecosystem-centric CI opens up more Unknown Unknowns.
Another one of my favourite ecosystem CI tools is Google Trends. If you are going to analyze organic search behavior, why not directly go to the source and exhaust it's value before you pay for a differen tool?
I love analyzing broad trends in intent using Google Trends. Here's one that I might be interested if I was running a food company, and wanted to know what to bet the farm on… Comparing user interest in tofu, seitan, tempeh and hummus…
Look at hummus go! It is the next salsa. :) On a serious note, this data is really good for looking at long term trends and identifying seasonal interest patterns. Try running a query with your industry name, and the top brand names in your space. That is how you get Share of Search, and amazing insights.
If I manufacture both tofu and hummus, I can use the Geo report to precisely understand where the indexed demand is coming from and adapt my marketing and advertising strategies accordingly. Currently my tofu ads are in heavy rotation in Pennsylvania, Colorado and Michigan. Oops…
Compare and contrast with hummus… A significantly more diverse consumption pattern…
Looks like everyone in America loves hummus. (I'm so surprised that California is missing from the top 7. That is the value of querying a database of intentions like Google Trends, you realize how wrong your gut is.)
Close out by understand the fastest rising queries for both….
Use this data for your SEO, PPC campaigns, to invest in creating a new flavor of hummus (can anyone confirm if cauliflower hummus is any good, it sounds a bit yucky), and even, as you can see above, celebrities you should associate with.
If you want to learn more about other amazing analysis you can do, checkout my post on Google Trends (under it's old name): Google Insights for Search: Competitive Analysis.
Raven Tools is another great ecosystem analysis tool, though they have evolved to be a lot more (including your single repository for all your Google Analytics, Social and other data). I've used their Research Central to do website and keyword research. Here's a sample report for a brand I utterly love and adore… Patagonia…
One of the really cool thing about Raven Tools is that it pulls data from many sources, Google, Majestic, Moz, Calais and others. This allows you to dig deep into data quite easily.
If you are purely interested in Paid Search effectiveness, the AdWords Keyword Planner is an amazing resource. Lots and lots of search behavior data laid bare for your consumption…
While it was built for PPC, you can easily see how you can extract numerous insights for a more effective SEO strategy. Primarily because your Google Analytics/Adobe/IBM data only shows you how much you got, here you can see how much traffic is available!
You can also drill down…
And glean a strong understanding of seasonality which is useful for every type of paid and organic campaign you want to run – including display and social.
Speaking of display… When it comes to finding where you can find audiences who might be most interested in your products or services, you can take advantage of the free Google Display Planner.
Continuing the fun we are having in ecosystem analysis, I give you Kim Kardashian!
You would do this for your own site, but what we are looking at is the analysis of where to run advertising if we were in-charge of www.people.com and our customers were interested in Kim Kardashian (#doh).
You can of course use these lovely recommendations to figure out where to optimally target your display ads (look at that mobile number!), but also to more broadly understand attributes of your audiences, the sites they tend to be at and so much more.
A final recommendation, one you might not think fits in this category.
Google Analytics has benchmarking reports for your key metrics. The GA team pulls together analysis from your industry vertical, country and like-sized websites. These end up as benchmarks from the ecosystem that are integrated into GA and available for you to extract insights from.
Here's how it looks like in your channel grouping report…
It is important to take advantage of this ecosystem CI data so that you can have stronger context to your performance. To learn, extensively, how to take advantage of this data, see the post.
I hope this gives you a feel for the breadth and depth of ecosystem-centric competitive intelligence analysis you can perform. With just a little bit of work, you can glean spectacular insights.
That's all of it. Everything I know, and love, about competitive intelligence analysis.
It is my fervent hope that you will fall in love with it too, if you don't already use this data. And if you do, I hope you learned some new tricks and new tools.
As always, it is your turn now.
Are you convinced that CI data can provide incredible context and insights to your site analytics data? Are you just a little bit less in love with Adobe/WebTrends/Google Analytics? For site-centric analysis, do you have an absolute favourite go-to report for acquisition or behavior? What insights have you found to be actionable in traffic reports? Is there valuable context you need for Search from CI tools that we've not covered here? We rushed through ecosystem-centric analysis, but is there one analysis you are now newly determined to perform?
Please share your advice, favorites, critique, and life lessons via comments below.
Thank you.
As always – a pleasure to read your posts, only managed to go through half post though but will be back later today to finish reading. CI Analysis is one area I love exploring more and this post centralised everything in one place.
Waiss
Lovely post! CI definitely is an area every marketer / analyst should be utterly interested in.
We all suffer from blind spots or tunnel vision sometimes, that's why it's important to get a new perspective on things. What are THEY doing?
To help us PPC and SEO people, I find SEMrush to be really helpful, even though it focuses more on the keyword side of things (aquisition) – it quickly lets you identify opportunities, holes in your keywords and competitors you didn't really have on the map.
Again, thanks for the great read Avinash.
Hi Avinash, thank you for the post, actually i was working on global competition for one of our client.
It's great to find information like this in right time. I agree: SimilarWeb it's not so accurate as it could be, but it's very helpful and transparent. Data is mismatching Analytics data of course but it's still nice comparison tool.
Thank you one more time for great post!
Best
Bartek
Bartek: Accuracy is a tough word to use in the world of CI, for any tool.
Each has its strengths and weaknesses, we just have to figure out which tool has the best strengths for our situation. : )
Regardless, in the real world we have so little to go by but in the digital world, thanks to even imperfect data, we have a lot more to go by. :
Avinash.
Thanks for the wonderful post, Avinash. You'd asked on Twitter if I had any tips to add. Here are a few little tips not covered above:
1. Monitor the cadence of your competitors' website changes. For example, if you're in small Ecommerce, check their websites once per day to see how frequently they update, when sale launches, etc. (there are off-the-shelf tools for doing this automatically gathering screengrabs; I made one to do this & categorise the screengrabs by brand/industry/etc & found it really useful). If you're working at a larger scale, it may mean more specific checks. Once you're doing this, you sometimes get to learn odd things. Eg, 'oh, they usually leave that feature spot live for at least a week, but they've pulled the jeans promo after just 2 days; it may not have been working', etc.
2. Same as above for email. Sign up to your competitors' emails. If you run regular surveys, ask the question "which of the following do you buy from?" including your competitors, to understand the overlap. Once you understand these 2 pieces of information, you can avoid sending out mails after theirs, avoid crossing categories, send out your sale email before theirs, etc.
3. A slightly odd one, but for ecommerce: placing test orders at regular frequencies from competitors *can* give you an idea of their sales volumes. Most ecommerce platforms simply use sequential numbers for orders. With that knowledge & a rough guess at average order value & return rate, you can reverse engineer a fairly good guess at their financials.
4. It's there in plain site, but most people don't notice it: Wikipedia give away all their stats. This can be very handy as they sometimes rank for competitive terms, and it can give you an idea of market change, and of general traffic volumes. To reach it, click 'view history' (top right on desktop on any page), then 'page view statistics' (again, toward the top right).
5. URL shorteners give away a lot of info, and most brands use them in social media channels. Simply add a '+' to the end of most shortened URLs (bitly, buffly, smarturl.it, etc) to gather some stats. On bitly in particular, if a user is logged in when shortening a link & has not set their profile to private, you can browse through their entire shortening history including stats.
Plenty more, but that's enough for now! Happy to add more if useful.
dan barker
Thanks for the info!!! And Avinash too, of course :-)
Great adds here Dan. Especially #4. Adding to your suggestion I sometimes look at stats.grok.se if you don't already know it.
Thanks!
Robert
Dan: I'm deeply appreciative for the time you took to add these thoughtful pointers. I'm confident, so many readers found them to be immensely helpful.
I agree if you are a small sized competitor, keeping an eye on their site changes and promo schedules can be immensely valuable. For larger sized competitors, I simply sign up for their emails. You learn so much about their marketing strategy. Different kind of CI data.
I have to admit I'm not so sure about #3 for most companies, the Order Id might be a tough thing to read too much into, and to go from there to AOV etc. is probably skating on thin ice.
With #3 I thought you were going to go recommend understanding fulfillment, and holiday impact on that. Perhaps there is something there?
Using the + to get short url stats is my fav. I'm amazed at how many people don't know that. So helpful in understanding social effectiveness of content and even time of posts (for ourselves and our competitors!). Thanks for adding this one.
Avinash.
First of all thanks for the amazing post!
I would just like to remind you all that up until about 2 years ago we've had a brilliant (and free) tool that Google has sadly (and against their mission to “organise the world's information and make it universally accessible and useful”) sunset: Google AdPlanner. I understand their reasons, but given the mission… :( Anyways.
And it's also worth reemphasizing that SimilarWeb sadly still doesn't have a good enough solution for mobile usage outside the US (and perhaps the UK).
Zoltan: I'd linked to the Google Display Planner in the post above, it is the evolved (and still free!) replacement of the Google AdPlanner. It does many of the same things, though a couple of my favorite bits are not there any more – and for those I share your disappointment.
But do play with the Display Planner, if you thought AdPlanner was good.
It would be great for everyone in this space to get their act together and start building out a bigger mobile dataset (or any mobile dataset!). Meanwhile, SimilarWeb at least has something.
Avinash.
Thank you Avinash.
As a fellow CI junkie, it is great to gather additional perspective on the message in the data. Love the additional points from Dan as we find great insight in competitive website creative and massage monitoring. a few additional thoughts…
We also find great value in understanding not only where competitors are buying display ad space but with what creative & message. For this we look to comScore's Ad Metrix and/or WhatRunsWhere. The same can be said for the value of ad creative on paid search and for this we use SpyFu's ad history report.
On the SEO front, we love our SEOMoz subscription. In addition to on site factors we find OpenSiteExplorer extremely valuable when evaluating competitive link building strategy. With some digging you can gain great insight into how the competition goes about acquiring new links (ex. content strategy) and with what keywords as a focus.
Finally. when assessing email strategy, in addition to joining the lists as suggesting by Dan we also look to tools like Hitwise for spikes indicting pushes and then monitoring volumes from webmail domains. when we map volumes from competitors programs against that for the programs we run, in some cases we can even make assumptions around list sizes.
Thank you again for a great post!
Jeff Rasp
Very interesting article.
Thanks for the time and the effort it takes to create something so valuable.
Hi Avinash,
I have used Follow.Net on a couple of occasions to get a sense of traffic reports + traffic sources and the % weren't too off. What are your thoughts on this freemium CI tool – Follow.Net?
I have used it in the past while working on agency pitches as it helps us understand direct traffic (brand awareness), get a sense of social traffic (provides more context than mentioning just fan base sizes) and helps us look at the current SEM/Banner ads on to critique the creative. It pulls charts from Compete, Alexa and SimplyMeasured and shows quite a bit of info. for free.
P.S. – I do not work at Follow :)
Thanks.
Adil: Follow seems to be aggregating data from different public sources, and collecting data from your browsing after you install their Chrome extension (data that they then use across multiple Follow users as a signal for their analytics – very much like Alexa and others). Dig a bit more about the latter, and make sure you are comfortable (some people are, some aren't).
I'm not quite sure what's unique about them, to switch away from current sources with bigger datasets that are also free to a certain degree at least. But I'll play with it more.
Thank you for sharing it.
Avinash.
Genius.
(I especially love the diatribe about 1/4 of the way in on why demographics don't matter. It's a point I have struggled to make with management who is in love with magical ideas of big data. But not the point of the article.)
Anyway, regarding the caveats are about how the data isn't there for little guys. My experience is that for B2B, even if you have some size, the data from folks like HitWise means little since business users generally don't use the big ISPs (the ones that send data to HitWise). And visit data past the few top-level pages was highly suspect. YMMV. Interested in hearing from others.
But all that said, there are insights in the data and thank you for showing how to sniff them out. Some clever ideas here.
Moe: You are absolutely right, B2B is also a challenge. Even if the click behavior can be collected, there is not enough of it.
The bigger problem, is that most competitive tools are now relying panel based on measurement, and most companies (certainly larger ones) will not allow any kind of intrusive monitoring software on their employee machines.
It is critical to be aware of this, and deeply quiz the vendors as to how they are accommodating for these issues.
Avinash.
Wow, that is the demon of all posts.
Great on insights of competitive analysis and how most of these tools are useless below a certain data threshold, for a lot of sites probably applies.
What got me was you tips for tools that help present the data better. That is the winner here as well. We are currently in a data overload and any means to be able to present data quickly and concisely is a winner.
Top quality as ever.
Awesome and purely explained, it's really worthy content for me.
I am possibly cutting my nose off to spite my face but I was immediately put off of reading this article when I saw the lack of English as an option on the adverts for your books. Normally that wouldn't bother me, but the addition of American as a language (considering that it doesn't actually exist in its own right) actually angered me enough to comment.
Otherwise, your introductory paragraph was alright.
JB64: I'm sorry, but I'm not sure I understand your comment, would you please share a bit more context?
You'll see that my books are published in many different languages, the information is in the right nav of each page on this blog. Those links are so on the book's website I've linked to in paragraph one in this post.
I hope this helps.
Avinash.
I think he means the American flag on the right side of your book adverts. He wants to see the English flag too.
John: Thanks so very much for helping clarify! I was really confused.
I respect JB64's right to be upset, and the desire to see the Union Jack there instead. I'm an American, I'm comfortable with the choice I made (though I have to genuinely admit I did not give it as much thought as one might imagine).
-Avinash.
Yep, John got it right. Apologies for derailing the thread, I think I got here in the wrong frame of mind. But still, whilst I think you'd have to be crazy to disagree that the USA has the dominant Western culture, I still take some pride in the fact that the most widely spoken language in the world originated from our little island.
Thank you both for your replies. I'll now continue reading the rest of the article…
Hi Avinash,
Very nice post about competitive analysis. Thank you for that.
A bit a pity that some data can just be seen with the paid version of the CI tools. Nice to have similarweb for some quick comparisons though.
Regularly reading your posts, always good tips. Thanks.
Roland: I agree, it is a bit sad.
But, for these tools, to get to the quality they should be at, they have to invest money and they don't have any other way to make money. Hence, they have to charge.
There have been many a startup that has launched CI tools, but they fail after a while because they can't figure out how to survive.
Avinash.
Fantastic post. I tweeted out a series of tweets (@deanabb) related to the behavioral vs. demographic data you commented on and I couldn't agree more:
I've been building predictive models for 27+ years and have dozens of opportunities to build models with both behavioral and demographic variables. I've given up using demographic variables when they compete with behavioral (if behavioral data is available) because they never win. Demographics help sometimes help augment data (in some interaction) but I've found that's even rare. Almost useless.
The reason for this is that usually the target variable in predictive models is, …, well, …, behavioral. Likelihood to purchase, attrite, lifetime value, etc. Demographics, as great as they are for reporting and understanding your customers, are usually too broad to make decisions at the individual level.
So now I've usually given up even trying demographics for predictive models even though I routinely hear folks in retail talking about "personalization", meaning demographics. Again, if you know nothing more, it's better than nothing. But if you have behavior, use it.
Dean: Amen!
While I recognize that both you and I are in a minority, I am happy to have a kindred spirit.
Sadly, it will take a while for the Sr. Leaders in control to unlearn a lifetime of one way of thinking, but I'm optimistic it will happen.
Oh, and even when the smart ones say personalization, they really mean understanding a person's behavior and then reacting. They just call it demographics! :)
Thank you so much for sharing your experience with all of us.
Avinash.
Aviniash,
This was a great article. I work in CI and I have a question for you about the differences in competitive tools.
I am comparing two websites using Hitwise, Compete, and SimilarWeb Pro.
Hitwise is showing that website A is killing website B for visits. Compete is showing that website A is beating website B for visits, but not as much as Hitwise. And then Similarweb Pro shows a different story that website B is beating website A for visits, but not Crushing website A.
My question is, who do I believe? Why would two CI tools show the same thing, and the other not? Similarweb Pro also has a huge panel of more than 10m people, larger than Hitwise I believe. Since the panel is larger, should I go with Similarweb Pro's findings?
If you could help me out, that'd be AWESOME!!!
I really enjoy all of your articles.
-Aleah
Aleah: As you noticed I did this comparison at the start of my post, and they all show different numbers, and that is ok.
Spend some time trying to understand each tool's strengths. Compete for example is purely US centric and has strength in that population. SimilarWeb's has global data, it might be thinner here or there. Quite ok of course because very often it is your best choice.
Hitwise has not evolved over the years, though they've intimated that some changes are under way. At at the moment you are better off between the two above.
Here's the coolest part. You can simply buy a month's membership, validate them, then close your subscription for the other one. Even at their full list prices, it is cost well worth bearing.
Avinash.
Hi Kaushik,
I read your post & its quite good. Do you know about any CI tools which can compare data between apps as majority of our traffic is coming through apps.
Thanks & Regards,
Anand
Anand: SimilarWeb has some mobile app data in the Pro version.
For mobile apps, the competitive intelligence tools ecosystem is still in the nascent stage. I'm sure this will change with time.
Avinash.
Thanks Avinash :)
Great article! In-fact I had implemented a few techniques before bumping into your post.
From my experience, when I presented my findings to the upper management, the first question they asked me was "How accurate is the data that you have?" and they actually want a number for it!
Its kind of a tough one to answer on the spot so I decided to do a little research myself. I compared the data from CI tools to the data from our web analytics tool that we have deployed for our website. The deviations for basic metrics is around 7% – 25%. The deviations are large since a single percentage change can lead to a difference of a couple of thousand visits.
So help me out here. How do I sell it to the management or atleast a way to show that the data being employed in the CI is fairly accurate?
Jagjit: The answer to your management team's question is: "The data can't be compared to our web analytics tool, or the competitor's web analytics tool. But. Trends in the data, and specific metrics for dimensions like traffic are of value."
That's just the way things are. I cover at the start of this post, why the data will never be "accurate". Each methodology has its own biases. There is simply no way around it. In fact, if you repeat your exercise in a few months, the same tools might have shifted!
If they simply want accuracy, there is no choice. They can't use CI data.
It is still worth trying to explain to them because data you get in CI tools you can't get anywhere else (including equivalent options in the real world).
Avinash.
Hey Avinash,
Thanks for responding. I did try my level best to push my findings. Seems like the digital field is new and hard for the old school marketeers to comprehend. I don't think the problem is with the data accuracy instead its a problem with the mindset of my management.
Anyways, I will keep pushing data backed intelligence within my team.
Cheers,
Jagjit
Wow, talk about epic content! This post is so long and value-packed I had to look through it a few times. Great stuff.
I can't believe I've never heard of SimilarWeb. Looks cool – I'm going to give it a shot now.
Thanks
These are all really good. This is a fast growing sector for Internet Marketers. I had never thought of using Google Trends like that before.
I personally use Mozenda to harvest data / scrape from the web in order to get information about my competitors. I had a company ask me for all the retail locations of one of their competitors and they were amazed with I came back the next day with the full list of 600. It is a cool tool.
I also like to use usertesting.com to learn a bit more about user experience.
Thanks for this post Avinash and the older one comparing Hitwise versus ComScore, very helpful as I'm just looking into purchasing some CI software for my company.
Do you have a view in terms of the paid version of SimilarWeb versus ComScore (Hitwise doesn't cover the international markets I am interested in)? ComScore gets a lot of mentions around the web as one of the big players but SimilarWeb seems to be a lot newer and it is harder to find feedback on it.
Considering the quite different price points, my concern is that by going with the cheaper option, it isn't as reliable.
Thanks.
Rums: I'd advised Aleah that one simple approach might be to simply buy access to the tools for a month, stress-test them, and then pick the one you like the most – the one that has most type of data/reports you need.
With regards to ComScore, I would encourage you to investigate how they collect their data, and their sample size outside a few big countries. The answers will be helpful in deciding if you want to use them. If you want to dig deeper, understanding the sampling bias might also be of value.
Avinash.
As always, great article.
I was a user of SimularWeb but must admit when Pro was launched in November I started to use RivalFox
But maybe now I will give it another try, your insight is as always an inspiration.
Thanks again
Paul: I'm so glad you found the post to be of value.
I've had the opportunity to use RivalFox (though there is no free version, they generously do provide a free 14 day trial). I'm afraid I find it a bit difficult to use, and in terms of data prefer some of the ones in this post.
Though, we can all, because of the free trial, use it and find the best fit for our use case.
Thank you!
Avinash.
Hi Avinash,
I'm a keen follower of your blog. I'd like to ask something based on your experience in web analytics:
How can we evaluate and possibly quantify the impact and performance of a web analyst in a company?
Christos: Thanks for this great question.
We would obsess more about the impact of the insights the Analyst has identified (rather than think of the Analyst in a silo).
This blog post shares a specific strategy you can follow to compute ROA:
~ Excellent Analytics Tip #22: Calculate Return On Analytics Investment!
The post contains identifying overall cost of analytics, and how to identify ROI on various digital activities. Be sure to download the free spreadsheet at the end to jumpstart your ROA efforts.
Avinash.
Thanks for giving this very useful information and sharing it.
It is rare for such valuable information to be freely available.
Thanks.
Hi Avinash,
Thank you for the great book Web Analytics 2.0 you wrote! I am reading about bounce rate and depth of visits in page 94. The figure 4.16 shows that 76% of the visits are 1 to 3 pages, and visits are actually number of visitors (because you also calculated percent of all visitors). I assume a report like this is generated for some period, like weekly or daily.
I don't understand how number of visitors are counted. If I were a visitor, I could get counted in multiple rows, because one time I only looked at one page and then I was gone. Next time I come, I went through 4 pages. So, I would appear in the 76% visitor list, as well as the remaining 24% visitor list. So, the visitor lists associated with each depth are not mutually exclusive. How would you suggest to deal with this situation? Would you remove me from the 76% list and keep me at the largest depth I appear? If I want to know who are the 24% and likewise the 76%, obviously the list of visitors needs to be mutually exclusive, isn't it true?
Number of visits and number of unique visitors obviously are not the same. But I see in this table, and your writing in that page, you tend to use them interchangeably.
KC Tung: You are absolutely right, the first part of the page is right when it emphasizes that the report shows the distribution of visits.
The sentence you refer to should read: "You can see that 76 percent of the Visits by website Visitors did not exhibit the ideal behavior." It should not read: "You can see that 76 percent the website Visitors did not exhibit the ideal behavior."
Avinash.
Hi Avinash,
Would you be so kind as to do a review of semrush.com? Your views on the strengths & weaknesses of their data (e.g. I don't think their ranking data includes mobile searches) would be interesting and useful. This for me (I'm in the UK) is the best competitive analysis tool.
Thanks for a great blog.
C
CB: I appreciate the kind invitation, but I'm afraid I'm unable to do reviews of tools.
I do hope though that there is enough guidance in this post about reports and metrics you should look for, and how to evaluate each tool (more on this in section one of the post).
All the best!
Avinash.
Hi Avinash,
Love all your blogposts thanks so much for sharing. Have you looked at competitor analytics using BuiltWith or Datanyze's data to see what software / analytics tools competitors are using?
Deb
Deborah: I have not, but I'll take a look.
I've generally cared very little about software or tools that competitors use, because it is rarely the software/tools that make a material impact. It is People, Process, Strategy. Sadly, don't quite have a tool to get competitive intelligence on those three! :)
Avinash.
I'm interested to learn which tool you used above that was showing you visitors' profiles and demographics, with the gender and ages, especially the one for below 18 (displayed in your image).
Manal: That screenshot is from Quantcast.
In addition to the other tools mentioned in this blog post that also provide demographic, psychographic data… you can also get it inside your Google Analytics account. More on that here:
Google Analytics: Demographics and Interests
Avinash.
Hi Avinash,
I enjoy reading your posts especially about how further you can go on a certain topic. The level of depth and how clear you present an approach is a great find for each data curious human. As I was researching about CI tools and the metrics around the web an idea came up and I want to suggest it here if its ok for you too. I haven't read anything in depth (your style of depth) about a very valuable competitive intelligence metric which we track so little.
Pricing metrics of our competitors.
It's one of the most important metric out there that is still working to outrank a competitor in the year 2018 either behind closed doors or publicly open in the web. It will still be the dominant factor for a sale in any democratic money-centric economy. And because of this, I'd enjoy reading a post about your take on the intelligence of pricing metrics because you have a different approach to any kind of analytics, data etc. so far.
I might be biased as I work in a pricing intelligence software and might be exaggerating this pricing metric a bit. But when I think of it over and over before writing this comment it really affects our decision-making B2C, B2B, doesn't matter from whom to who.
So, what are your thoughts about competitive pricing intelligence as a metric for any type of company, person etc.?
Yigit: Pricing if of course an important bit of competitive advantage (/disadvantage). I have not covered it here because it falls outside the influence of what Digital Marketing typically does.
Let me add that while price is important, and in the commodity business especially it is the key determining factor, I do not believe that price is the be all and end all. I often end up paying more for a product on Amazon simply because of the convenience and the peace of mind that if something goes wrong Amazon will make it right (they always do!). I also end up paying a lot more for my light down jacket at Patagonia because I deeply love the mission of the brand (it is irrational but when I wear my blue Nano Puff I feel better!).
Chasing price can be a race to the bottom without the role of price in our broader business strategy.
Avinash.