The usual suspects you might forget in your campaign reporting
If you are working in the Digital Marketing Analytics domain, you might have been confronted with the request: “Can you tell me how the campaign XYZ performed?” Of course this sounds in your ears like a business-as-usual question… and of course you have an -ideally automated- way to come up with the right set of KPIs and pack them into a more or less sexy report.The campaign landing page had an average number of visitors of X, an average conversion rate of Y.
The total number of sales attributed to the campaign is Z. (and a long etcetera you know very well)
So far, so good… but let me invite you to keep reading… and you might find out that something might be missing in your analysis… Let me run you through a few points people usually just neglect in their campaign performance reporting:
1.- The baseline or the definition of “normal”
To understand the real impact of the campaign, we need to put all KPIs in the overall performance context. What works best in this case is defining a norm or a baseline at different levels to compare with.
The baseline shall be as unbiased as possible but at the same time, close enough to the campaign to enable a proper comparison.
For example: let’s say you are running a campaign to sell the new Motorola Nexus 6 bundled with a large data plan. Your baseline shall consider the performance of all campaigns advertising a high-end smartphone with a large data plan as well. Considering all campaigns ever run say for prepaid plans with low-end devices or focused on a tablet might compromise the referenceability.
Once you have a baseline, you can make statements like: “our campaign is improving the conversion rate of the standard high-end device/plan campaign by 1.8 percentage points”.
Moreover, having a baseline allows for setting upfront the proper campaign goals and steering according to these goals: “The expected click through rate (CRT) for a Display campaign is 1.2%, whereas our current campaign is still at 0,98%”
2.- It’s too hot to buy anything… Understanding seasonality
Purchasing behavior over time shall also be considered. Let’s say you run the same campaign -same product, same message, same marketing spendings- in summer and then in December. Obviously you expect to sell more in the pre-Christmas time than in the middle of August. Making your performance relative to the natural seasonality shape is a mandatory step for a proper campaign assessment.
It requires a set of modeling decisions, but it’s worth it and necessary… and if done properly, supports the upfront goals setting mentioned in the previous point
3.- You are not alone – Understanding the competition
I find really astonishing, how companies still think they are alone in the market. To me, this point of view is equivalent to calling the prospect customers something along the lines of “stupid”… not able to use internet to compare offers, products and services. Obviously this attitude is not the best guaranty to survive in the market. When you run a campaign, you are obviously competing with existing campaigns from your competitors. No matter how fiercely you are pushing your products, your success is subject to the offers and campaigns of your competitors.
Incorporating this information to your data set and “annotating” your results accordingly allows for better understanding what’s really going on with the campaign. What also helps is defining a competitiveness index for your offer over time, which encapsulates the competition situation.
4.- Jumping on the bandwagon is not necessarily bad
Porter did include as player in his 5 forces model, the suppliers as one of the main players. You might found that one of your suppliers is running a campaign already for a product on its own, that you can capitalize on. Best example is Apple and the Telco operators… There was a time, where no marketing effort was required by the Telcos to sell out the entire iPhone stock within few days… Apple -the supplier- was advertising it and it was more than sufficient. Obviously this is an extreme case, but the learning here is to couple your campaign with an existing campaign with the same target. Understanding what your suppliers are doing is therefore part of the exercise of assessing the performance of a campaign.
5.- The perfect wave
Nobody is really free from hypes. You need to understand the shape of the demand for your product/offer. Think of a surfer waiting for the perfect wave… If you jump too early, you are going to struggle to come back afloat… if you jump too late, you are going to lack traction and sink after a few seconds. So synchronizing demand cycles and campaigns is a must. Measuring the demand cycles is also a must to assess the performance of a campaign. Tools like google trends can help here.
6.- Yes, there’s an offline world
Fortunately, not everything takes place online. There’s still (intelligent) life outside internet… and these living beings watch TV, read newspapers, check daily their (physical) mailbox, etc. Understanding what has been done over these channels is also essential to assess the performance of a campaign. Let’s say again you are advertising the new Nexus 6 and you regularly broadcast a neat TV ad… You can expect a decent volume of searches in the form of “[your brand name] Nexus 6″… which you wouldn’t have if you didn’t 🙂
This assistance needs modeling and interpreting… yet is very often forgotten! If you have ever tried to understand or to build yourself a ROPO (Research Online Purchase Offline or the other way around), you know it is everything but simple… but without that, you can be certain your multi-channel strategy is lacking fundament.
My take on that
I’m with you it’s not easy to even get the data assets to allow this kind of campaign assessment… and come on! you already have your way of reporting campaign performance… and I agree, you can do a decent job reporting the hard KPIs (visits, visitors, sales, conversion rate, etc)… but unless you understand the market pulse, the context and the competition you are not going to know if the campaign was really successful… and without this information you are in a really bad position to optimize anything.