Are you confused about what parts of your digital marketing are working? I suspect you are and if you aren’t hopefully today I do a little bit to make you less comfortable about your certainty. No, this isn’t some cruel April Fool’s joke – it’s me wanting to be helpful. While sewing confusion might not seem to be helping anyone, I’m hoping what follows gets you to ask more questions and to refocus your efforts a little. I’ll add, as any good teacher does, that if you’re still confused come see me (OK, call or email me) after class for extra help.
Source: farm2.static.flickr.com
Let’s start with a quick story from my TV days. When the college football overnight ratings would come in they would be one number. The overnights were 25 metered markets, mostly the biggest 25. When the national ratings came in a few days later, the ratings would have changed. One might expect this as the rest of the US was now included. However, when you’d look at the Northeast region, for example, the rating might be very different even though nearly all of the population was included in the overnight ratings. We’d get told it was two different samples which, of course, were measuring the same thing in the same area but through different methodologies and different homes. It was extremely frustrating.
Fast forward. We’re deluged in numbers. The problem is that many of them measure the same thing but give us different answers. Take search. You want to know how people search for your site or product. Google Analytics is mostly useless now since Google’s (not provided) result tells you nothing and represents a ton of your search traffic. Webmaster Tools provide some search term information but when you compare some of the other information with the same data points in Analytics the results are shockingly different. Which do you take as gospel? Add to that the data you get from AdWords – also different – and you’re now thoroughly confused.
Speaking of ads, most of the clients I know look at the top of the conversion funnel – how many people saw an ad. The problem is that some studies say 50%+ of ads are not viewable. Obviously that affects conversion rates, ad copy effectiveness measures, etc. You also have these kinds of issues with content publishing on other social platforms and broad measures such as “likes” and “follows.” The social guys are doing a better job of cleaning up fake accounts but there is still a long way to go. The results of a content campaign shown to 5% of your followers that are real vs. 5% that are fake will obviously vary widely.
What can you do? First, look more at trends than at any data point and second work backwards. Metrics such as sales (lower-funnel metrics) are hard to get wrong. Each step back up the funnel increases the uncertainty somewhat so be wary and ask questions. Experiment, watch trends, measure sales, rinse repeat. Just be careful about attributing that success to anything based on measurement tools that might have fogged up in the heat of battle. You can’t see very well though lenses that are mostly obscured.
OK?
Consult Keith 