Troy Lerner is president of Booyah Advertising. Follow him on Twitter @troylerner.
All attribution vendors are not created equal. I’ve done a lot of hands-on research about various attribution solutions available. As I’ve become more comfortable with the tools and methodologies available, a few things have become evident that every marketer should consider when selecting an attribution product.
You have read the maximum number of free articles.
This content is available exclusively to Digiday+ members.
Products that offer “full-funnel attribution” yield significantly better insights and are worth the investment. Completeness of the underlying data set is the only way to ensure accurate attribution results and recommendations. Having the ability to confirm the quality of the data set is also essential.
And it matters how the attribution vendor assigns credit to the conversion. Currently, most providers use one of two different methodologies to assign credit. The more accurate approach assigns credit based on analyzing actual campaign data, which allows you to make objective investment decisions. A simpler, but less accurate, technique relies on pre-determined assumptions on which campaign attributes deserve more credit. Making decisions based on arbitrary assumptions like these does not produce actionable results.
Note that data sets are far too complex and large for analysts to manually evaluate and optimize. And by the time you find any helpful correlations, it’s likely too late to do anything meaningful with the information. Rather than having to re-build and re-run models multiple times, a sophisticated attribution tool should uncover as many insights as possible in an automated fashion. And the outputs from these insights need to be automated so they can feed search engines, display platforms and other media-buying venues, allowing tactical recommendations to become practical to implement.
No optimization tool is flexible enough to handle all of the campaign constraints that a client might have. There are contractual, as well as policy reasons, that live outside of the data that need to be respected. That’s why it is critical to have robust modeling tools you can use to adjust recommendations based on varying constraints. Once the media buyer has the core optimization recommendations, she can use what-if modeling tools to apply constraints.
Finally, there are also a few things you should actively avoid when searching for an
attribution partner. For example, be skeptical of any vendor that promises immediate
returns or “out of the box” insights. The truth is, there’s no such thing as one-size-fits-all attribution, and for every campaign and client, it takes some level of customization and understanding of business goals in order to make attribution work. Steer clear of products that promise results without investing the time to understand the business and its goals.
Similarly, avoid real-time attribution reports. Several vendors tout products that deliver
real-time results. I hope we get there someday, but I don’t believe it yet. Actionable
attribution analysis typically requires several weeks or even months of data.
Finally, make sure your contact includes access to the nerds. Getting high-quality attribution insights is complicated and requires lots of data and a flawless trafficking/tagging system. You’ll need to put your own talented folks onto the project, but you’ll also need regular contact with the experts from your selected tool provider. The sales guy will certainly promise you this access, but it’s worth asking the technical experts directly if this is their understanding as well.
Attribution tools have matured greatly, but many vendors still sell solutions that sound (and are!) too good to be true. Keep in mind that you’ll likely be engaged with your attribution partner for a long time. It’s a significant investment of time, effort and money, so pick a tool that serves your needs and a provider that will stand by you. You’ll be glad you did.
Image via Shutterstock
Sign up to get the day’s top stories at 6am eastern.