Dan Robins, Head of Interactive at OMD Sydney, examines exactly why we should stop neglecting attribution, and how we should be implementing it into our customer engagement approach.
Attribution is one of these things a lot of people know should be done but feels just a bit too hard, a bit too complex to get started. So you put it in the too hard basket. Once it’s there, it seems more and more difficult.
It’s been neglected for a long time, more so in Australia than in the rest of the world. In Australia, we’ve been a bit slow to take up the technology available to us. The market is developing, and there are a few local brands doing attribution and doing it well, but for the majority, it’s been a bit neglected. We’re slowly getting to the realisation, “We do need to do this, and it’s actually not as difficult to get started as we might have thought.”
We know our customers see and interact with our ads across all media channels as a whole, not in silos.
A lot of brands are wondering how to get started. It’s so complex. Then how do you go from starting to a fully built out, multi-touch, bespoke model? The answer to that is have a go. Get started. You can have an eye on the end goal, but simply comparing different touch points within one channel can be a great start. Using the out of the box models in your ad server or web analytics is one starting point. Any testing is better than no testing.
The brands taking on the attribution challenge
There are quite a few cases where brands have tackled attribution and shown the value to the overall bottom line. We look after attribution for a number of our clients here at OMD. We’ve got an in-house attribution team and our own technology working for brands across our group. Other clients have selected specialist agencies to run their attribution model. What they all have in common is an approach of starting small and testing over a period of time.
While starting is important, it’s equally important to be prepared for the long haul. Attribution is the sort of thing you need to commit to it. It’s not a golden bullet. You can’t just do it and say, “That’s ticked, done now,” and it’ll live on forever more. You need to continuously be working with it, trailing, optimising, and trying again. It’s a big commitment and when there are so many other opportunities for brands to consider such as trailing programmatic or other campaign ideas, it’s not as sexy.
It’s a difficult nut to crack to do it long term, but we’re certainly starting to see some of our clients test the waters. It’s a case of generating some data with the tools available, make some assumptions based on science, trial and optimise from there.
Who is responsible for attribution?
One of the other issues is the proliferation of specialist agencies working with brands. A brand could have a specialist performance agency, a specialist trading desk, perhaps a specialist digital agency or specialist search agency as well as its wider media buying agency.
To pull all of those together to have one overarching attribution model is difficult. It takes quite a strong brand manager or CMO to say, “This is what we’re doing,” rather than having to corral multiple agencies.
The ultimate responsibility of attribution comes down to the brand. It’s their responsibility because it’s their budgets being managed. Whoever the lead media agency is, it’s their responsibility to be neutral and transparent over the decisions and optimisations being made, as well as to manage the process openly and fairly. But there’s also an onus on any specialist vendors to engage. Poor engagement will lead to inertia.
So who do you have as the overall owner? To me, that’s whoever your lead agency is. That’s the main thing, but moreover, who do you trust the most to be focused on your brand’s best interests?
Which attribution model is right for my brand?
For brands, often the biggest struggle to overcome is the range of attribution models and approaches available.
Instead of attributing all sales to the last click or view, out of the box models can be set up to assign a value to the first click, share data across points in a linear fashion or in time decay. For the so inclined, it’s also possible to adopt a U-shaped model which assigns a high value at the start and end of the customer’s path to purchase with lower values in the middle.
In terms of the actual technology, off the shelf models can be pulled out of most ad servers and many tech stacks. That’s probably the first step. Ultimately, however, the technology needs to be bespoke to the advertiser and driven by how customers interact with the brand. The type of advertising a brand utilises will also have an impact.
It is important to mix the science with some media smarts. Blindly following a model’s output to switch funds between channels without direction on what is actually viable from that channel and its diminishing return, is a recipe for disaster.
The next step for attribution is across devices: being able to map how consumers interact with your brand on a desktop computer as well as on their mobile or tablet. The technology isn’t quite there yet but it’s certainly starting to catch up.
Beyond that, the goal is being able to match a consumer’s interaction with your brand on mobile with their interaction in the real world, taking into account whether they went past a particular out-of-home site, saw a TV spot at home or using their mobile as the identifier. It is a bit of a leap, but it’s on the horizon now rather than being just a figment of the imagination.
Of course, the first step is getting brands to look at attributing touch points across digital interactions because if they’re not doing that, it’s going to be a pretty big leap to track those non-digital interactions.
This article originally appeared on Ashton Media.