Tips for Data Driven Decision Making in Marketing

20th October 2015 0 Comments

Having the data is only a part of the puzzle. Marrying that data with your inventory of available next actions and ideas is the really tough part. Here are some quick rules of thumb for informing and executing better decisions using data within your marketing activities.

  • Budget for test, and then test, fail, recover quickly, and then test again. People are naturally resistant to failing, but with new data from everywhere available you, failure is nothing to be afraid of. Experiment with new marketing actions using your data. If you have a hunch, use some budget and test for it. Failing fast is not a bad thing. It should be encouraged, because it means you are discovering what does not work for your business while pushing you out of your comfort zone and challenging your assumptions. Knowledge of what doesn’t work is equally valuable than knowing what you have always done. In particular, consider A/B testing: keeping all but one external factor the same, change the design or flow or function or conversion target for some marketing collateral you have and then measure the results. One recent post by 37 Signals shows the millions in lost revenue associated with a homepage redesign that was not A/B tested. 
  • Understand that correlation does not necessarily mean causation. This is one of those lines you remember from a statistics and algebra class in university but you don’t always remember what it means. Essentially, the takeaway is that just because two things seem related does not mean that the same event caused them to be that way. One universally accepted example of this is a study that links bigger shoe sizes with increased reading aptitude. You might be inclined to conclude from that data that having bigger feet means you will be able to read more quickly and with better accuracy, but of course that isn’t the case. There are other factors at work.
  • Segment your data based on behaviors or types so that you can derive relevant insights from each group. It is easy to see this massive blob of data and think that all of your customers are represented equally within it; however, nothing could be further from the truth. Different segments of your demographics have different data sets associated with them. Depending on how you acquire customers and leads, by what channel they come in, based on what conversion action and value they are assigned, the data will lead to separate insights for each of the separate segments. Take the time to properly segment your data so that you are teasing the most accurate, quality data and informing your marketing decisions with the highest integrity.
  • Consider the sources of your data. Unfortunately some data is more accurate than other data. This can be for a variety of legitimate reasons: sometimes there is capture error, in that some data can’t be read from some customers or prospects. Sometimes there is a transmission error where data is either duplicated by mistake or not sent at all also by mistake. Sometimes there is an analytics error where the data processed by your analytics platform might be using the wrong corpus of records. Some data from web analytics is often wrong or misleading—there is a whole market of people, for instance, who can tell you what Google Analytics is right about and wrong about when it comes to measuring web traffic data. Here is one great post that speaks to that phenomenon. Ultimately, it is incumbent upon you to be careful to make good decisions using two or more sources of data that make sense against each other. Avoid where possible making decisions (unless it is on a test basis) based on known questionable sources of data or single sources of data, or where you are not able to reconcile conflicting records from two different sources of data.
  • When all else fails, ride the trend lines. The direction values are moving are much more important than specific numbers, for all of the reasons I mentioned in the previous paragraph. Rarely will you get down to the single visitor or penny or shilling the same number on all your platforms. Rather, look at trends—are visitors increasing? Are conversions improving? Are more customers dropping out of your sales funnel at step D than before? Are visits down? The trends will tell you all you need to know to inform better marketing actions; don’t waste your time trying to swim in absolute numbers.



Share This