It turns out that Marketers, especially Digital Marketers, make really silly mistakes when it comes to data.
Big data. Small data. Any data.
In the last couple months I've spent a lot of time with senior level marketers on three different continents. Some of them are quite successful, but sadly many of them were not. In the latter group I discovered two common themes:
[A benchmark for you: If you are not spending 30% of your time in 2013 with data, Ms./Mr. Marketer, you'll fail to achieve professional success.]
2. Many used some data, but they unfortunately used silly data strategies/metrics.
Silly not in their eyes, silly in my eyes. And silly simply because as soon as the strategy/success metric being obsessed about was mentioned, it was clear they would fail.
A silly metric, I better define it :), is one that distracts you from focusing on business investments that lead to bottom-line impact. They get you fired. Sometimes sooner. Other times a little later than sooner.
I'll expand my purview a little bit in this post, from just looking at silly metrics to also looking at silly data myths. My hope is to use these eight examples to illustrate to you how, if you are spending 30% of your time with data, you can use crazy cool data strategies/metrics to ensure many, many promotions.
Eight data myths that marketing people believe that get them fired:
2. All you need to do is fix the bounce rate.
3. Number of Likes represents social awesomeness.
4. # 1 Search Results Ranking = SEO Success.
5. REDUCE MY CPC! REDUCE MY CPC NOW!!
6. Page views. Give me more page views, more and more and more!
7. Impressions. Go, get me some impressions stat!
8. Demographics and psychographics. That is all I need! Don't care for intent!
Excited? I am, let's go!
A lot of people get fired for this.
Sadly not right away, because it takes time to realize how spectacular of a waste of money getting to real-time data was.
It seems absolutely stupid to say, "No, I don't want real-time data." It's like saying no to Chocolate/Jesus. You just don't say no!
But I want you to.
I want you to say: "I don't want real-time data, I want right-time data. Let's understand the speed of decision making in our company. If we make real-time decisions, let's get real-time data. If we make decisions over two days, let's go with that data cycle. If it take ten days to make a decision to change bids on our PPC campaigns, let's go with that data cycle."
Here's why… Real-time data is very expensive. It is expensive from a systems/platforms/data processing/data reporting perspective. You end up paying substantial amounts of money to your analytics/big data vendor – for questionable value. It is also very expensive from a decision-making perspective because if you have real-time data you'll darn well make sure that it is being shoved down every single person's throat. That approach will ensure that, even in the best case scenario of the proverbial pigs flying, they'll obsess about tactical things. All the time. The normal case, after a few days of "OMG I can't believe I have data! It is real-time! Call my mom!!!" their eyes will glaze over and they will ignore all the numbers flying by because they won't know what to do with it.
So shoot for right-time data.
That is a cheaper systems/platform/data strategy. (And remember even the most idiotic system in the world now gives you data that is a couple hours old with zero extra investment from you. So when you say real time you are really saying "Nope, two hours is not enough for me!").
That is also a way to get people to sync the data analysis (not data puking, sorry I meant data reporting) with the speed at which the company actually makes decisions (data > analyst > manager > director > VP > question back to manager > yells at the analyst > back to director> VP = 6 days).
As a result of the above two outcomes, you'll get promoted.
PS: The phrase "real-time data analysis" is an oxymoron.
PPS: I've mentioned one exception in the past. Real-time data is super valuable if zero human beings are involved from data collection to action being taken. Say for example, artificially intelligent platforms that do automated trading on Wall Street.
No. You'll get fired. In three months.
The difference between a KPI and a metric is that the former has a direct line of sight to your bottom-line, while the latter is helpful in diagnosing tactical challenges.
Bounce rate is really useful for finding things you suck at. Bad ad creatives. Horrible landing pages. Poorly targeted ads. Missing calls to action. Et. Al. You measure bounce rate and you can find those things, then figure out if the problem is at the source (ads) or destination (your site). Finally, like magnificent little troopers that you are, you'll fix the problems.
Along the way you also learn how not to stink. Bounce rate goes from 70% to a manageable 30%. Takes three months.
Stop obsessing about bounce rate. By lowering your bounce rate all you managed to accomplish is get your ads created and targeted properly and optimize the landing pages. You got the person to the site, they did not puke and leave right away. Great. Time to focus on rest of the process.
From the time people land on your site it might take another 12 – 25 pages for them to buy or submit a lead. Focus on all that stuff. The tough stuff. Then you'll make money.
Fixing the bounce rate is mandatory, without that you are coming to play the super bowl naked – you are not going to win.
But just fixing that won't ensure you win. Focus on the actual game. Focus on incredible behavior metrics like Pages/Visit, focus on the Visitor Flow report, obsess about Checkout Abandonment Rate, make love to Average Order Size.
All that will get you promoted. Because you are going to focus, not on metrics, but on key performance indicators that have a direct line to the bottom-line of your company.
Ohh … this one will get you fired even faster. :) Because it does not take a very long time for your Senior Management to figure out how lame the Likes metric is and that it drives 1. Zero value on Facebook and 2. Zero squared economic value or cost savings to the business.
There are many spectacular reasons for why Like (and +1s, Followers) is a horrible metric. Here's one: We are looking at two consumer product brands, the tiny company Innocent Drinks and the Goliath called Tide Detergent.
Checkout the number of Likes each has.
Innocent has 0.3 million and Tide has 3.7 million. Small difference.
Now if Tide was lame they would use the number of Likes they have and boast about that to their CMO. Bad idea.
Tide is not lame. They also look at the number right next to the Likes number. For Tide that number reads 15k. For Innocent that number reads 58k!
Even with 10x the number of Likes on Facebook the giant called Tide has 4x fewer people talking about their brand when compared to the David called Innocent.
That's primarily because Tide does not truly understand how to do Conversation Marketing. Their relentless day after day mix of "here are pictures of our products repeated every day, here are the people we sponsor and their ads" is insufficient to create brand engagement.
And if Tide were lame and used Likes as victory, they would not know this.
On the other hand, Innocent clearly understands Conversational Marketing (they are a small company, not yet corporatized I suspect, and likely deal with a smaller combination of lawyers/PR types/Social Media Gurus/Agencies – a sad but likely bane on Tide's existence). They manage to get hundreds upon hundreds of comments (Conversation Rate) and thousands upon thousands of comment Likes (Applause Rate). I'm not using hundreds and thousands as metaphors:
[Update: As no less than three comments mention below, Innocent is 90% owned by Coca Cola. Fooled me! Please read the above section in that context. But I have to admit this only makes me even more impressed with Innocent. In a massively large company they've carved out an identity uniquely their own. They refuse to be corrupted by Coca Cola's own Facebook strategy of constant self-pimping and product ads masquerading as "updates." As a result pound for pound Innocent's fan engagement on its page is multiple time better than Coca Cola's - even if the latter has many more likes. So Innocent, I'm sorry of thinking if you were a David but I'm mighty proud that you can remain that inside a Corporate Machine. There is hope yet for all other brands at P&G, Unilever, Nestle, Kraft, Pepsico, J&J, etc. etc.]
Tide is not lame. They are not going to obsess about the number of Likes. They will measure the four sexy magnificent best social media metrics and use them to drive a magnificent conversation marketing strategy on Facebook (and YouTube and Google+ and Twitter and Pintrest and …).
Neither should you Ms./Mr. Marketer. Because Likes (and +1s, Followers) measure a fleeting "hello." It is what you do after that first hello that creates business value. Focus on that, and you'll get big 5. Buying targeted impressions = business impact.
Not going to happen.
So what is it? When you search for a brand or a category term, most companies, small or big, want to show up number one in the search results.
Much money is spent on all kinds of demands to SEOs, big or small, which results in all kinds of shenanigans to try and get to #1. Sadly, all that yields very little.
The main reason, as all decent SEOs will tell you, is that search results are no longer standardized. Rather they are personalized. I might even say, hyper-personalized. Regardless of if you are logged in or not.
When I search for "avinash" on Google I might rank #1 in the search results because I'm logged into my Google account, the engine has my search history, my computer IP address, it also has searches by others in my vicinity, local stories right now, and so many other signals.
But when you search for "avinash" your first search result might be a unicorn. Because the search engine has determined that the perfect search result for you for the keyword avinash is a unicorn.
[Do the search wherever you are, do you see me or a unicorn? Share via comments!]
You should give up obsessing about only the number one keyword ranking for just this reason. But of course there are many others. Universal search for example means that personalized results will not only look for information from web pages, they also look for YouTube/Vimeo videos, social listings, images of course, and so on and so forth. Then let's not forget that, proportionally, there are very few head searches and your long tail searches will be huge. Oh, and remember that searchers rarely type just a word or two, and instead use long phrases.
There are a ton more reasons why obsessing about the rank of a handful of words on the search engine results page (SERP) is a very poor decision.
So check your keyword ranking if it pleases you. If your boss insists that your brand must rank #1 for x, y, z keywords, before you show him/her progress log, into Bing on his/her computer, search for x, y, z and click on your brand (even if it is on page 2), and chances are when he/she searches for x, y, z they might see your brand ranking #1. :)
But don't make it your KPI.
There are many other better, no-way-I'll-get-fired SEO metrics.
For purely SEO, you can use Crawl Rate/Depth, Inbound Links (just good ones) and growth (or lack thereof) in your target key phrases as decent starting points. You can graduate to looking at search traffic by site content or types of content (it's a great signal your SEO is working). Measuring Visits and Conversions in aggregate first and then segmented by keywords (or even key word clusters) will get you on the path to showing real impact. That gives you short-term acquisition quality, you can then move to long-term quality by focusing on metrics like lifetime value.
Don't start with LTV, it is hard. Use the step-by-step process above. And you'll get promoted.
I. Hate. This. One.
H. A. T. E.
It upsets me greatly when I see companies create entire Pay Per Click / Search Engine Marketing strategies based on this single metric.
I've heard this more times than I care to remember: "Our CPC for our Bing/Yandex campaigns was $2.25. This quarter the goal of our PPC campaign is to get that to $2."
This is the image that comes to my mind.
Image Credit Frits Ahlefeldt
Let me use an example here.
Let's say you have a stock portfolio with someone like E-Trade. As a prudent investor you'll buy low and sell high. (Easier said than done!) When you trade stocks you'll incur a per-trade fee from E-Trade, usually $10.
Would you measure the success of your trades based on cost per trade? Would your overall trading strategy be to reduce that price to $9 next month?
Of course not, that would be insane.
You will measure the impact on the bottom-line as a result of the trades. How much profit did you generate?
Likewise for the results of your Bing/Yandex campaigns.
You should not care if the CPC is $2 or $25. If someone is on Bing and raising their hand that they want to buy a vacuum cleaner, then you should be willing to pay to show up to sell a vacuum cleaner because that is what your company does.
You should judge the success of that showing up by measuring whether you made money! Did you earn any profit?
As you buy paid search tools, hire smart people/agencies, you should measure if you are able to make more and more profit. You should measure whether, over a long period of time, you are able to increase the lifetime value of PPC acquired customers.
CPC is a profoundly silly indication of your search strategy success. It causes companies to make silly decisions, which leads to poor results (both for the digital strategy and your career).
Friends don't let friends use CPC as a KPI. Unless said friends want to get their friend fired.
Content consumption is a horrible metric. It incentivizes sub-optimal behavior in your employees/agencies.
If you are a news site, you can get millions of page views writing stories like (not kidding, all real headlines) "10 Reasons Apple's Innovation Machine is Dead" or "The Kardashians Are On Vacation In Greece" or "Guy Performs 'Do My Thing' In Sailor Moon Outfits Made Of Construction Paper." And it will probably get you transient traffic.
Then that traffic goes away. So your next story is "22 Longer Sex Condom Ads From Around The World."
That works for 40 minutes. Then what?
And what about business impact from all these one-night stands ? Unless the spammy display ads on your site are getting you $14,000 CPM, your company won't be able to afford your salary for too long.
If you are in the content-only business (say my beloved New York Times) a better metric to focus on is Visitor Loyalty (and if like me you are a paying subscriber: Customer Loyalty). Rather than focusing on a transient metric, you focus on your site's ability to deliver such value to the visitor that they come back again and again – and pay you!
If you are in the lead generation business and do the "OMG, let's publish an infographic on dancing monkey tricks which will get us a billion page views, even though we have nothing to do with dancing or monkeys or tricks" thing, measure success on the number of leads received and not how "viral" the infographic went and how many reshares it got on Twitter.
If you are in the ecommerce business, the only reason to care about pageviews (in a pages/visit context) is to obsess about understanding how to create the most optimal shopping and purchase experience on your site (and the fewest pageviews to happen in the checkout!).
Don't obsess about page views. Spend a short amount of time thinking what causes your salary to get paid, the bottom-line. Then measure the metric closest to that. Hopefully some ideas above will help get you promoted.
Display advertising is an integral part of any digital marketer's repertoire. As it should be.
But I'm heart broken when the success of their Facebook/AOL/Doubleclick/Video ads is purely based on impressions.
If you are forced to watch, say a TV or Digital Video ad, perhaps we can summarize that you actually saw it. You can imagine how tenuous the connection to impact is for "impressions" of banner ads (all shapes, sizes, levels of intrusiveness). Did they really see them?
My hypothesis is that TV/Radio/Magazines have created this bad habit. We can measure so little, almost next to nothing, that we've brought our immensely shaky GRP metric from TV to digital. Here it's called impressions.
Don't buy impressions.
I know you are holding a gun to my head so I'll say this instead: Buy engagement. Of course define what it means first.
I'll be a smidgen more delighted if you evolve to measuring clicks on those ads. It is a signal that someone saw something and clicked on it.
If you are willing to go to clicks, do one better and measure Visits. At least they showed up on your mobile/desktop site.
Now if you are a newbie, measure bounce rate. If you have a tiny amount of experience, measure Visit Duration. If you are a pro, measure Revenue. If you are an Analysis Ninja, measure Profit.
Impressions suck. Profit rocks.
Perhaps you disagree. That is ok. We are both reasonable people. You can buy impressions if you can prove via a simple controlled experiment that when we show impressions we get more engagement/sales and when we don't show impressions we do not get more engagement/sales. And I'll be extra sweet to you, the display/video impressions don't even have to be last click! Actually you don't even have to have a click!!
If the simple A/B (test/control) experiment demonstrates that delivering display banner ad impressions to the test group delivers increased revenue, buy impressions to your heart's content. I'll only recommend that you repeat the experiment once a quarter. See, I can be reasonable.
But if you won't do the experiment and you use the number of impressions as a measure of success, don't get too comfortable in your chair. Bye, bye, coming soon.
This is not a metric, this is more of a "what data you'll use to target your advertising" issue.
It’s another legacy from the old advertising channels like TV, magazines and radio. Our primary method of buying advertising and marketing is: "I would like to reach 90-year-old grandmas who love knitting; what TV channel should I advertise on?" Or they might say: "I would like to reach 18-to-24-year-olds with college education who supported Barack Obama for president." An example of demographic and psychographic segments.
We would have no idea if Grandma actually wanted what we sold or if the young person was remotely interested in our service/product. But we had no intent. We just had results of questions asked during dinner by phone surveys and content habits. We took that on very thin ice data, we bought advertising. That was our lot in life.
[Did you know 50% of TV viewership is on networks that each have <1% share? Per industry.bnet.com. I dare you to imagine how difficult it is to measure who they are, and how to target them to pimp your shampoo, car, cement.]
We do the same on the web. We go to Google+ or Facebook or AOL and say "Hey, you have a billion people, give me their age and education and let me send a deluge of display ads to them or push my promoted posts!"
The click-through rates prove the poorness of that strategy. Just look at your own reports.
Intent beats demographics and psychographics. Always.
So if you have advertising money to spend, first spend it all on advertising that provides you intent data. (Don't worry, you'll have plenty left over to gamble on demographic/psychographic data. I promise.)
Search has a ton of strong intent. It does not matter if you are a grandma or an 18-year-old. If you are on Baidu and you search for the HTC One, you are expressing strong intent. Second, content consumption has built-in intent. If I'm reading lots of articles about how to get pregnant, you could show me an ad related to that (even if I'm a man, I'm actively expressing interest in getting pregnant!).
The first intent is strong, the second one is weaker. (If you use remarketing you can make that second intent even stronger.) But in both cases the most important thing is not how old I am or my gender or my education or the number of legs I have.
There is a lot of intent data on the web. That is our key strength. Don't waste your marketing and advertising on mental models for targeting, delivery and engagement that were created in the age of Mad Men. Leverage the mental models of our day and age and you'll be immensely successful.
I wish you all the very best!
Ok, as always, it is your turn now.
There is perhaps no limit to the number of silly mistakes Marketers (or pall bearers or butchers or …) can make. Do you have favorite mistakes that are not listed above? Which personal "mistake" taught you the most valuable lesson of your career? Do you agree with the eight above? Which one do you most disagree with? Any thoughts on the cultural challenges that cause us to make these mistakes?
Your insights, advice, helpful tips and critique are most welcome.