Many analysts are talking about big data and multi-attribution platforms. Attribution is great for providing some insight into how campaigns may be overlapping and provides some summary level reporting on your primary conversion. However, if you really want to measure how a campaign is performing, you need measure that campaign against its goal.
Attribution platforms are just that, platforms. They are built to tell you all of the many marketing touches a user encounters before converting; which ones deserve credit and how much credit each deserve. Basic attribution models allow you to look at every touch point and pass credit evenly among them. With more advanced models, you may look at common paths to conversion and compare them. One path has touches A, B, and C, and the other has A and C to determine the impact of touch point B. This is fairly simple when done once, but attribution platforms are making this kind of comparison hundreds and thousands of times depending on data volume.
The impact of the advanced systems is that your conversion attribution data is in a state of constant change. As every conversion comes in, the value of each touch point in that conversion is determined and the impact of any marketing can change all past conversion values.
While the advanced systems may make reporting a challenge because it is regularly changing, it is still the route many practitioners are taking. I agree with the approach and often will use the attribution data for both reporting and for analysis. However, attribution data may not tell the whole story, although it is often portrayed to do so. Instead, the recommended route is to measure each campaign for its specific intent.
Each marketing campaign should have an intended impact on the users who interact with the campaign. If you are running a display campaign, the intended impact may be to increase brand awareness. If you are running a retargeting campaign, the intended impact may be to bring a potential customer back to the site. Each campaign has a distinctly different intended impact, and therefore should be measured differently. If the display campaign generates a large number of impressions, and a recent survey shows an increase in brand awareness, did the campaign fail if people did not convert? No, the campaign did exactly what it was supposed to, but without a conversion, a multi attribution system is not going to give much indication of the success of the display campaign. Instead, it is important to measure each campaign for its intent.
Moving Beyond Basic Campaign Measurement
Beyond the success measurement of each individual campaign, it is important to get a big picture understanding of how you are performing. Yes, attribution data can be helpful here. However, using a traditional funnel is also a great approach. Measure your awareness campaigns for ability to drive awareness, your interest campaigns to drive interest, your desire campaigns ability to generate desire, and your direct response campaigns ability to elicit action. When you are measuring your branding campaigns ability to generate a direct response you are setting the campaign up for failure.
Using your marketing funnel and campaign goals, you can setup a funnel overlay that will show you the success of each campaign to hit their specific goals, and the impact of top level campaign performance on lower level campaigns. Analyzing how each campaign can drive users down the funnel will give you a path to determine budget allocation on a consumer path to conversion level. Within each step in the path to conversion, the specific campaigns can be measured against each other to optimize performance. As you observe paths to conversion outside of the determined path, for example, users skipping steps, you can start to re-allocate budget to the appropriate steps, removing allocation in this example from the step users are skipping. This process is highly dynamic, as you shift budgets and start to see the funnels shrink and expand, you will be constantly optimizing and adjusted towards the ideal ratios within the funnel.