Why real-time attribution is (mostly) irrelevant, and what to do instead
The martech and ad tech industries love to talk about real-time. The context doesn’t matter — regardless of the circumstances, real-time must be better than delayed. Right?
The allure of real-time makes sense. We operate in real time these days, at work and at home, answering calls and emails as they come in, fighting fires as they arise, watching our customers and prospects browse websites and click our offers literally as it happens. The real-time perspective is certainly compelling and sometimes valuable from an operational point of view — but not always.
In fact, when it comes to marketing attribution, real-time results are far less important than right-time insights.
Forrester’s Tina Moffett articulated the concept of right-time insights in one of our meetings. It immediately resonated as an intuitive description of all the hype around real-time and something we had already worked into our product.
The truth is that most media doesn’t work in real time. Every channel works at a different pace and cadence of effectiveness. The length of time in market for each channel before it drives behavior varies dramatically. There are applications and channels in which real-time measurement makes sense (programmatic, largely), but the majority of the time, you have to give media some time to do its work.
A focus on right-time attribution changes the conversation from “how quickly can we get results” to “how do we get the insights and data to the channel at the right time, so that we/it can make a better optimization decision?”
When is the right time?
Again, that right time will differ for each channel. For programmatic, it does demand close to real-time input. Channels like video, and especially TV, on the other hand, perform better when optimized based on daily, weekly, biweekly or even monthly insights.
You need to give media in these channels a chance to make an impact. If you optimize too quickly, it simply won’t work. Just because a TV spot didn’t drive action today, you shouldn’t necessarily optimize it out.
Instead, by determining the right maturation cadence for each channel, you gain the ability to optimize strategically. That may mean moving money from publisher to publisher or changing contracts — decisions that require more time and thought than shifting budget from keyword to keyword.
In some cases, real-time isn’t even realistic. With a TV campaign, for instance, you may have 50 percent of your budget caught in an upfront. Even if you had real-time data to push to the network, what are you going to do with it? A right-time perspective enables you to work within your buying constraints.
The right time will depend on your industry as well. While there are some truisms when it comes to channel cadence — TV moves slowest, programmatic moves fastest — your optimization strategy should also reflect your specific business. The cadence for retail and commerce, for example, is much shorter than automotive, regardless of channel, and those differences must be taken into account.
Of course, the right time for some channels will be real time. Programmatic can and should support real-time insights. In a bidded search program aimed at the bottom of the funnel, constant bid optimization at the keyword level absolutely makes sense. Feeding information back into the engine in real time or near-real time makes the process smarter and allows it to optimize more effectively — and that’s a good thing.
Your attribution solution needs to be able to support real-time, too. The real-time use cases are indisputably important. But programmatic shouldn’t be the tail that wags the dog. Strategic optimization across your media ecosystem requires adjusting your cadence to each element of the mix, not always catering to the lowest (or fastest) common denominator. Your attribution tool must be in sync with this strategy and able to deliver to the requirements of every channel.
Putting right-time optimization to work
What does right-time optimization look like in practice? Pushing data into all media decisioning environments so that you can make the best choices for that channel.
In a programmatic environment, as described above, algorithms can quickly consume that data and act accordingly — great. But you can use attribution data to inform all sorts of decisions, not just automatic algorithmic ones. The best results often come from humans and machines working together with democratized data, processed at the right time and with accurate context.
Real-time optimization isn’t wrong, but it is a limited view of how the real world works. To get the most from your data and your attribution program, consider right-time, channel-specific insights instead.