Attribution 1.0 – five lessons learned

Alison Lohse runs through five lessons learned from Attribution 1.0 and how they apply to today’s technology.

Alison Lohse 180Every new technology experiences growing pains, and marketing attribution is no different. While the idea of attributing a conversion to a specific action, whether in a store or on the web, is as old as business itself, sophisticated solutions for measuring the interplay and impact of those actions are relatively new. Attribution 1.0 was a huge step forward from last-click measurement. But it also left a lot to be desired. First-gen attribution technology was, in a nutshell, a one-size-fits-all solution. Each attribution vendor’s approach was rigid; once you signed up, you were locked into their methodology. As a veteran of the space, I can attest that this our-way-or-the-highway approach was not malicious. It was classic 1.0 software: a great idea that wasn’t fully baked.

Fortunately, the market has evolved, and we’ve learned from our mistakes. Let’s break down five lessons learned from attribution 1.0 and how they apply to today’s technology.

1. Getting up and running shouldn’t take forever

In the old days – and even today for some vendors – an attribution platform could take up to a year to fully implement. Because each vendor had a fixed methodology, the client was forced to adhere to that model, appending tags, changing operational processes, pulling ad server logs – the list was long. It was also expensive, and in the end, not particularly effective. Modern attribution tools adapt to your circumstances. The pace of insights matches the pace of media, the market and optimization needs. The technology also maps to your existing infrastructure and operations to retrieve and process data without compromising its fidelity. That means faster, more accurate results without the never-ending timelines and expenses of a rigid tool.

2. The more algorithms, the better

The original attribution platforms had good reasons for their rigidity: they operated off a single, proprietary algorithm. That single algorithm dictated that data could only be construed one way, regardless of a client’s particular scenario. We’ve now discovered that using multiple algorithms creates more precise outputs that are more relevant to the customer. One way to approach this is the ensemble method. Analogous to the “wisdom of crowds,” an ensemble framework blends results from multiple models to achieve a prediction accuracy that’s higher than any single model on its own.

3. The fewer reports, the better

OK, not completely. But attribution 1.0 tools fire-hosed users with all the data, all the time. Outputs weren’t customized, and users had no influence over what information they saw or how it was presented. As a result, many users experienced acute information overload. The data was there, but it was often obscured, overlooked or misinterpreted. Next-gen attribution platforms surface what’s most important. Users can create custom reports within the tool, or bring data into other reporting systems. The data is liberated, which means it’s actionable: The most meaningful information is easily accessible, but it’s all there if you need it.

4. Being agnostic matters

Many early (and existing) attribution solutions were born or bought by the media organizations that they claimed to measure. While those vendors claim otherwise, I find that approach to be an inherent conflict of interest, and so do many of their customers. Media-agnostic platforms ensure the integrity of your data. With no alternate agenda or technological preference, an agnostic solution operates from a neutral point of view and adapts to whichever media partners the user requires.

5. Humans matter more

Don’t get me wrong; technology is incredible. The business insights gained from these intricate mathematical maneuvers give marketers more power than ever before. But first-gen attribution tools delivered predictive analytics that also presumed to be answers. Vendors sold their solutions on the merits of cutting out the media practitioner middlemen with reports that contained the whole truth and nothing but, and could be handed straight to the CMO.

Not so fast. The best attribution systems don’t replace humans, they empower them. By providing practical insights in a digestible format, next-gen attribution platforms arm media practitioners with the knowledge they need to make informed – human – decisions. There is still automation. There are still reports. But in the end, the machine adapts to the human, not the other way around. Marketing attribution technology has made remarkable strides in my 10-plus years in the industry. As technologists, I believe we have, too. By incorporating the lessons learned from attribution 1.0, we’ve created solutions that integrate, adapt and evolve to give users the most meaningful data yet. As a user, you just get to take advantage of them.


Alison Lohse is co-founder and chief operating officer at Conversion Logic.