How attribution modeling can help you maximize your marketing spend

Ian Harris

Executive Chairman

Digital Marketing Strategy

If there’s one thing almost all marketers have struggled with in recent years, it’s attribution modeling.

Attribution modeling is the science of assigning credit to the marketing investments that led to your customer making a purchase.

According to a recent Google survey, it’s the #1 tactic that occupies marketers’ minds.

As many marketers will testify, capturing, collating and interpreting the increasingly complex data sets about how, when and why customers make a purchase is not easy.

Customers might engage with your marketing multiple times, through viewing online ads, engaging with social media, and using online searches. They can do these things using a mix of laptops, PCs and mobile devices. And then they might walk into your store and make a purchase.

Even where the data is captured effectively, many attribution solutions struggle to generate meaningful insights as they use arbitrary, rigid models that don’t reflect the complexities and variables in your customers’ decision-making journeys.

As a result, few marketers can pinpoint the real value of a channel, and they continue to rely on guesswork about where best to invest their marketing budget.

In 2018, this should change, as more advanced attribution modeling technology becomes readily available.

Data-driven attribution models

The alternative to formula-based attribution models is to use a data-driven model, such as the model which Google has had built into its enterprise-level tool, Google Analytics 360, for some time.

Data-driven attribution works by analyzing customer journeys – both converting and non-converting – measuring how different journeys correlate with different conversion rates, and attributing value based on the impact each channel has on conversion rates.

For example, if Google was analyzing the value of your display advertising, it will compare the conversion rates of journeys that didn’t include a display ad with those that did. It might find that you have a 2% conversion rate across all journeys, but 3% when a display ad is involved.

Using this data, Google Analytics 360 can tell you how much value is added by each channel, a method that’s far more sophisticated than most standard attribution models.

Google Analytics 360 demonstrates this by comparing, side-by-side, its own analysis with a last-click model. The differences in the value assigned to each channel can be huge – Search Laboratory has seen discrepancies as large as 270%.

Data-driven attribution isn’t just built into Google Analytics 360. DoubleClick uses it as well, and everyone has access to the version that is available for free in Google AdWords, although this only includes advertising spend through PPC and the Google Display Network (GDN).

The latest developments in AI, however, and the escalating sources of data, have led Google to further develop this approach into an even more sophisticated solution that is due to be released some time in 2018.

Video: why now is the time to upgrade your enterprise attribution modelWatch now

Coming soon – advanced data-driven modeling in Google Attribution and Attribution 360

Google will soon launch two separate attribution modeling tools that use advanced AI techniques to build more sophisticated data-driven attribution insights.

The products will be called Google Attribution, which will be available free of charge, and an enterprise-level version called Google Attribution 360.

These tools utilize more data inputs than existing offerings, and they use advanced AI to analyze them from multiple perspectives to derive more statistically valid comparisons across channels.

Attribution 360 also includes forecasting and modeling tools. This means you can see how many more conversions you’re likely to generate, directly and indirectly, by increasing spend on a particular channel. When making these calculations, Google considers your diminishing returns curves, as well as the budgets of your competitors.

Used in the right way, these tools could allow digital marketers to make informed spend decisions based on an incredibly detailed understanding of ROI.

How to get ready for Google Attribution

Google’s new attribution tools have the potential to be game-changers, but marketers will need to put certain things in place before taking advantage of either the free version or the enterprise version.

Here’s what can be done to prepare:

  1. Review your Google Analytics (GA) account to make sure that the data you get from GA is correct and relevant, the value calculations won’t be very valuable. If you’re not 100% confident in your GA set-up, get it audited.
  2. Optimize your event tags. To make the most of the data you get from Google Attribution, you need to track the events that offer the most value to your business. You can’t get historical data once you set up new event tracking, so it’s better to sort this out sooner rather than later.
  3. Get excited.

It won’t happen overnight, and harnessing the data won’t necessarily be easy, but 2018 looks like the year that attribution modeling comes of age.

The days of arbitrary and ill-fitting attribution models should, at long last, be ending. We’re moving into an exciting new era for scientific, data-driven digital marketing.

Watch this video if you want to hear:

  • How to ensure that your attribution model contributes to your bottom line
  • How the recent shift in technology makes it easier to track your customers’ touchpoints and assign an accurate value to each
  • Concrete examples of data-driven attribution at work and what you can do to prepare your data
  • The latest updates on the Google Attribution tools

Need help getting your data in order?

This blog is part of our wider B2B Playbook that is designed to help B2B businesses with all aspects of their digital marketing from leveraging data, acquiring more traffic, creating assets that resonate and succeeding internationally.

In the data section we’ve compiled useful insights on everything you’ll need to know, from capturing data to building advanced machine learning solutions using your first-party data.

View all of our B2B data sources