Oscar’s seen a lot of changes in digital media in his years as a senior digital media manager. Here he discusses one recent trend that has had a profound impact on how digital can affect real-world results.
How did you originally get into digital media?
I always knew that I wanted to do advertising, even from school. I originally wanted to get into traditional media, but my first job out of uni was a digital role, as an account executive at a digital agency focused on web development.
I quickly realised I loved the hands-on aspect of being able to optimise results through learning. Something you can’t do in traditional media but can in digital media is instantly see a result improve after optimisation. It might be changing a bid on a campaign or tweaking audience or the creative. For me, seeing results change dramatically after just a few days is quite satisfying.
Is there an example of that happening recently?
Cisco Networking Academy is a good example. The client wanted to get more women enroling in its courses, so we conducted a small test using Facebook link ads, segmenting a female-only audience rather than letting Facebook optimise.
We tested for one month using a small portion of the budget on the female-only audience, and noticed positive results. So the next month, we implemented it at scale, and achieved a 70% increase in female enrolments with a 28% decrease in cost per enrolment.
What’s your main focus at krunch.co now?
My main focus now is around digital media strategy and campaign optimisation.
So, looking at certain media placements and their performance?
It’s important that we move away from analysing a placement and instead analyse an audience and recommend placements based on where that audience might be at a place and time.
It sounds a little like you’re talking about multi-touch attribution…
Not quite, but sort of. Attribution is a methodology – it’s not a new tool, or a new platform, it’s just a way of thinking about how we recommend placements and how we optimise the media mix.
It’s assigning a weighting to touchpoints that a user interacts with before doing a specific action – whether that’s buying a car or signing up for an event.
There are so many different models out there that we can use, however, the idea is to create a custom model that is catered to a specific business’s customer’s journey.
For example, the path to purchase a car is significantly longer than the path to purchase something like a beauty product online. We know that different channels have different impact in each customer journey, so when designing a media strategy, we look at those unique factors and the differing levels of intent on each channel and also each step.
What’s the point of multi-touch attribution?
The real point of attribution for businesses is to be able to see what placements are worthwhile spending money on (and also what’s not). It’s also valuable in confirming the effectiveness of campaigns and validating spend in certain areas.
But it’s important to actually use the data to keep improving. There’s a quote – “attribution without action is just a vanity exercise.” If you’re only using attribution to validate yourself, you’re missing a huge opportunity.
Also, there was a 2017 survey (by Econsultancy and AdRoll) in which 70% of respondents said their main frustration is that they’re not actioning the insights they’re getting from measuring the multi-touch attribution that they’re doing.
That’s due to an element of distrust – you’ve got a new way of thinking, and a lack of robustness, but the key to fix this is to create tests and keep learning. So we’re testing and learning constantly – trying different models, looking at different weightings… and we’ve got the ability to measure offline sales, so we can see a correlative change in digital when an offline sale happens.
What sort of potential does multi-touch attribution have for our clients and other businesses like them?
This is a somewhat new way of thinking, just over 5 years old. So there’s a lot of room to grow, especially around emerging technology that will help us even more.
For example, we’re working toward getting better visualisation on a real-time basis, and the ability to optimise in real-time to an offline sale or conversion, across multiple channels.
What is the most interesting thing you and the Digital team have managed to do with multi-touch attribution so far?
Just being able to link online activity to any offline event is pretty cool – for example, we can actually see and prove that a Facebook lead ad or Google banner influenced someone to book a test drive online, or even walk straight into a dealership without filling in any digital form beforehand.