Every Analysis Ninja knows that standard reports are lame. The lameness stems from the fact that they are created for everyone, stuffed with as much "stuff" as anyone might want and hence in the end satiate no one.
Custom reports on the other hand are, well, hand crafted by you for a specific purpose with a set of guiding principles ("Acquisition, Behavior, Outcomes! ") that ensure that they don't so much deliver data as much as deliver insights.
In this post I want to share with you three insightful custom reports for effective paid search (PPC) analysis. These insightful reports are from the six I'd created for my latest video for the Market Motive Web Analytics Master Certification course . [Checkout the detailed course description by clicking on the Curriculum tab.]
One of the hallmarks of our course (my course!) is that the content is constantly refreshed. Every few quarters we (I!) throw away 30% of the content and add more current knowledge. Makes the course a bit of a pain for the creators (me!), but substantially more relevant and useful for the students.
Rewiring the brain Teaching is hard work. :)
My latest video was on how to find meaningful insights, faster, when performing paid search analysis with web analytics tools (rather than, say, the AdWords or Bing UI). My emphasis was on getting Analysts to move outside the paid search silo, look more broadly at search engine marketing, and to not stop at keywords but rather dive into astonishingly important data elements like Ad Groups, Matched Query Types and more.
If you use Google Analytics then you can download these reports directly into your account and starting having immediate sexy data fun. If you use Baidu Tongji or SiteCatalyst or WebTrends or other Digital Analytics tools you can learn about the recommended analysis and recreate these reports inside your tool using its custom reporting feature.
The first part of my Market Motive video covers the best way to identify optimal starting points for improving bottom-line impact right away (I call it campaigns, keywords, outcomes) and how to find opportunities for making immediate fixes (optimizing ad copy, landing pages, match types ). After those critical steps we dive a little deeper…
#1: Optimization Lessons: Paid Search vs. Organic Search
One of the biggest mistakes perpetuated by siloed execution of SEO and PPC strategies is that the groups that own each piece rarely learn from the other party's work. Criminal behavior.
Search Engine Optimization will have a large source of your search traffic (surely you are not just renting traffic from Bing, Google, Yandex, right?). You can learn so much from bounce rates on SEO landing pages, the Per Visit Goal Value (or even better $index value) delivered, layout of the content, effective calls to action and so much more from your SEO dataset.
Likewise the secret gift of Paid Search is that you get to control the user experience, rather than the search engine. You decide what the keyword is. You decide what the ad copy is (your promise to the user). You decide the landing page experience. There are a ton of lessons you can learn from this controlled process that you can then turn around and implement on the rest of your web experience (and your SEO efforts!).
So why don't more people do it? Some if it has to do with company organization structure and incentives.
But there is also a tool reason. In Google Analytics specifically it is GA's annoying UI / user experience.
Paid Search is in a silo under Advertising (where nothing else exists, all other advertising you do is conveniently hidden away under, get this, Traffic Sources > Sources > Campaigns!). Organic Search is in hidden successfully under Traffic Sources > Sources > Search (where once again you'll see Paid in a silo but with different reports and options than under AdWords in the Advertising section!). The Overview report is in the Traffic Sources section and it shows a "dead end" breakdown of paid and organic. Oh and don't click on Search Engine Optimization, that has nothing to do with your keyword data (it is valuable but different, webmaster tools data, but does have keywords).
You can only compare paid with organic if you know a secret: Click on the small text called Keyword on the sixth of nine layers of nested horizontal navigational elements in Overview report. #omg
But what Google Analytics taketh away (the ability to do sane straight forward analysis), Google Analytics giveth (custom reporting).
Here's the custom report that will start your day off right… no data puking or fragmented silos, just the data you need to rock a little more everyday…
(Click on the image for a higher resolution version, or here: Paid Organic Search Performance)
Two lines at the top that show you quickly the balance between paid and organic search. This company could use some improvement but overall not bad. What you don't want is for the blue line to be really close to the orange, or across it. :)
You get immense focus in the scorecard (summary) using just the Acquisition (Visits, Unique Visitors), Behavior (Bounce Rate, Pageviews – proxy for content consumption) and Outcome (Transactions, Average Value, Revenue) metrics and Key Performance Indicators.
Underneath that performance for your individual keywords, again in an easy to understand layout.
You can see how this type of report can quickly show issues and opportunities.
If in your case you see that the PPC bounce rate is higher than that of organic visits then you should be sad. You are paying for this traffic, how dare it leave your site at a higher rate! In the above screenshot it is amazing that the PPC campaigns are a fraction of the traffic and yet greater than half of the transactions! For your first two keywords you can see that paid is indeed performing better than organic. So what can we learn from those landing pages? Go visit them. Look at the referring keywords. Why do they have a bounce rate twenty points, or more, lower than the organic landing pages? Why are the conversions so strong? What are the paid people buying compared to organic? What was the ad copy in AdWords, and is that distinct from the content on our organic landing pages? What else can we learn from the Visitor Flow report about content consumption (pageviews) for the two segments of traffic? What are the implications on improving the site user experience? What are the matched user queries for the better performing AdWords keywords, and are those actual user queries the ones you've done search engine optimization for?
Incredible questions, insightful answers for which, from your analysis, can have a immediate and material impact on your business.
As you dive deeper into the data you'll find places were your PPC campaigns are stinking, and organic rocks…
Similar questions. Find answers and make even faster improvements to your campaigns because remember every paid search visit to your site is costing you money!
Fun right? It really is.
Here's what the simple, but gorgeous, custom report looks like…
(Click on the image for a higher resolution version, or here: Paid vs. Organic Analysis)
A couple of things to note when trying to awesomize your custom reports:
1. Never create a custom report without Acquisition, Behavior, Outcome metrics. Each element gives context to the others and saves you from embarrassing jumping to conclusions.
2. Notice the drill down to Landing Pages from Keyword. Always try to figure out the most direct path to insights, and then build those dimensions directly into your custom report. And be brave enough to know when to stop. If two levels are all you need, two is all you should keep.
3. Leverage the built in Filters, they are god sent. Notice the critical choices I'm making above.
When you create custom reports using the Keyword dimension, Google Analytics sub-optimally does not realize you are doing search analysis and tries to get all traffic (!). I don't want that (and neither do you), so I filter out "not set." Less annoyance directly leads to a longer life span.
Second, since October 2011 searches made by users who are logged into their Google account you are going "not provided" and the keyword. We are trying to learn from performance of known keywords, so in this unique case I choose to filter out "not provided" as well. That might be an important chunk of your traffic, but you can't learn from people you know nothing about. In this case.
Be ruthless when you create custom reports. Be insanely relevant (pick the critical few metrics). Be insanely focused (filters, drilldowns).
Ready to do this with your own data? It is a simple two step process…
Save the report to the appropriate GA profile.
When the report is open choose the following standard built in advanced segments and click Apply…
Boom! You're in business.
Do your happy dance. Call your mom. Tell her how cool this is. Get back to work and answer questions. Make your mom proud!
#2: Campaign Optimization: Matched Search Queries Analysis
Now let me, gently, try to "unteach" you a small part of what I'd recommended you do immediately above. :)
When you log into AdWords and add Keywords to your Ad Groups, along with key elements like the price you're willing to pay, you also select your Match Type. Depending on your choice of Broad, Phrase or Exact your bids will be matched to relevant queries users type into the search engine.
So technically speaking your ads don't show up for your keywords, they show up for "Matched Search Queries." The words typed by the users. Yet most Marketers/Analysts rarely focus on them.
Here's a great illustration of this. We'd bid on the keyword "android notebook," and because of the way we'd structured our campaigns you can see that the actual user queries that our ads were matched with were quite varied…
Hence real optimization of your campaigns can't be achieved by merely doing keyword analysis of your AdWords campaigns. It will come from understanding what user queries those keywords/ads are getting matched for.
For example if "android notebook" gets matched to "android tablet" (most likely via broad match) then is that ok? Do we really want our ads to match people looking specifically for Acer notebooks if we have nothing to do with Acer? If not then perhaps we want to add a negative keyword (Acer) to our campaign to focus our money and ads.
That's just one example.
I like looking at the keyword reports as a great starting point. I love looking at the actual user queries because they reveal intent that is often hidden in the keyword report. Besides intent I also look for optimization possibilities to fine tune my ad groups because that will determine if my ads are shown by Google. And of course I'm looking for negative keywords to focus my AdWords.
To accomplish my analytical goals, I skip the standard matched search queries report as the hierarchy is not optimal and the metrics I want to report are more unique.
Here's what my custom report looks like…
You can see that per best practice I've chosen my critical few Acquisition, Behavior, Outcome metrics. Feel free to change these as you see fit in your Google Analytics / SiteCatalyst front end.
Then I start at the highest level and gradually drill down to the exact data I want. Source is the search engine (Google, Bing etc). Then the keywords I've bid on. Then the match type and finally which user queries are my ads being matched for.
You will notice that again to overcome the "here's a puke of all the sources in the universe" issue I've added a filter to my report. I'm using a regular expression to match values for Medium for just paid search campaigns. You can see the complete regular expression when you download the report into your account (below).
Here's how the report looks like, after I've drilled down two levels to identify the performance of matched search queries for one important head keyword I'm bidding on…
(Click on the image for a higher resolution version, or here: Matched Search Query Drilldown)
The keyword is calico toys and you can see a whole swath of user queries my ads are showing up for. Some, like Toys R Us, I probably don't want to match, others I might want to double down on.
The view you see above is just for Broad Match. The sweet thing about this custom report is that now you can go and look at what is happening to your Exact Match user queries (0.00% conversion rate! OMG! That is astonishingly rare for Exact Match!!). And then drill down to Phrase Match. In this case after my analysis we learned from what the matched user queries were for Broad Match and fine tuned that, and turned up the heat on Phrase Match (it should do much much better than it is currently).
Notice I'm not using Revenue or Average Order Size, rather I'm using Per Visit Value. I like using PVV, because I want to be a slave driver (sometimes :)), to drive the team to understand better how much money we can extract from each visit. AOV only tells you about each person who converted, which is good, but I like pulling the lens a bit higher.
Best Practice: While I love Per Visit Value, I really looooove using Per Visit Goal Value. PVGV tells me the business is focused on maximizing the value of every pay per click campaign visitor by measuring the economic value of macro plus micro conversions. That is not just winning, that is winning to the power of ten!
Ready to do this analysis? It really is super cool and super impactful.
Like keywords, love matched user search queries!
#3: Marrying AdWords Performance with Site Performance: End-to-End View
The universal challenge of web analytics is that our entire existence is subsumed by everything that happens after the click shows up on our website. Rarely do we care about what happens upstream. Often because we don't have access to that data, or sometimes because there are no incentives for us to care.
If your company is spending money (display advertising, YouTube campaigns, Bing, AOL, email campaigns, and Facebook ads) then you should go the extra mile, regardless of your job description, and try to understand what's happening upstream that is causing the clicks/visits to show up. Remember every one of those visits is costing you money. Yes increase revenue for those campaigns, but also obsess about reducing the cost.
For AdWords our life is made infinitesimally easier because by linking your AdWords account to your Analytics account rich AdWords data shows up automagically allowing you to have an end-to-end view of campaign performance. So lovely.
Here's my go to custom report, drilled down to the second level, for end-to-end AdWords analysis…
(Click on the image for a higher resolution version, or here: Paid Search End to End Analysis)
Pink highlights are the metrics showing AdWords performance, in green are the metrics showing onsite performance. Eight million (!) impressions result in 66k visits (what is up with that!). The average cost per click is 52 cents, the average revenue per click is 58 cents. Obvious question: Are we actually making any money on this? Remember the product/service you are selling still costs you money. What happens to the bottom-line, profit, on a six cent cost to revenue gap?
The sorting above is by RPC (where am I making the most amount of money, and how much is each of that click costing me ). It is hugely heartbreaking when you sort this report by CPC. After you wipe off your tears (and hide the reports from the HiPPOs) you'll log into AdWords to start optimizing the losers.
My other secret agenda with this particular report is to encourage you to understand the AdWords account structure that drives so much of your paid search advertising. Yes keywords are important, but in AdWords you set up Campaigns, inside each Campaign you set up Ad Groups which contain Keywords and ads that the themed for those keywords.
So when you do analysis of your AdWords campaigns an effective strategy is to set up the custom report to mimic the above structure. In my custom report I've done that by having an optimal dimension drilldown structure…
(Click on the image for a higher resolution version, or here: Search Campaign End-to-End Custom Report)
As always we have our Acquisition, Behavior and Outcome metrics (from the search campaign and our website). The dimensions then allow us to drill down from Campaigns to Ad Group to Keywords to the Match Type. We are once again using regex for Medium to filter and report on only the PPC campaigns.
I'm not kidding when I say that you could spend a lifetime understanding everything that is happening here and optimizing your campaigns. Though typically as a Web Analyst your job might be to identify valuable insights from this report and funnel them to your dedicated Paid Search team. They'll take it from there and do detailed analysis using deeper data available inside the AdWords UI (or your custom PPC bid management tool) and then take appropriate action.
Where you can really really help your Paid Search team is to put all this data in the broader site context and in context of all other non-search campaigns. They rarely look at that, you can then be your company's knight in shining armor by riding in on your white steed and rescue the Prince.
Ready for your happily ever after?
So there you have it… three custom reports that are focused on just what you need to do, are a delight to use, and help identify insights you can action to deliver an impact on the company's bottom-line.
Bonus: If you are a member of Market Motive or a student in the current Web Analytics Master Certification course then please watch the video on the site to download an additional three awesome custom reports for paid search analysis.
As always, it's your turn now.
How deep are your paid search analysis efforts? Does your company use the web analytics tool, or a specialized search platform for analysis? Have you use the metrics and dimensions used in the three reports above? Got ideas for how I can improve my custom reports? Or, better, do you have a favorite custom report that you use to analyze your pay per click campaigns?
Please share your experience, critique, examples, ideas and feedback via comments.
Couple other related posts you might find interesting: