#MustReadMonday: 2014 -The year mobile location data finally moved in the right direction

For this weeks #MustReadMonday, the very mobile Mark Stubbs – Account Director in Display at MEC London – talks about the maturing of mobile location data  and how MEC UK are demonstrating accountability  in real world behaviour. 

DATA

  • 2014 saw brands use location data in a more sophisticated way
  • Observing consumers mobile devices in-store has more uses than just serving an ad at that time
  • MEC are now able to track the cost per incremental customer for mobile ads driving consumers in-store

Nothing in media made me more frustrated or excited in 2014 than mobile location data. I tend to pick at least one side project each year to get under the skin of, and more than I care to remember don’t deliver -or haven’t yet delivered- on some fairly obvious potential (I’m looking at you TV ad synch…).

So whilst the time spent looking at mobile location data led to some late nights and a few heated debates, our work has been a success. At the start of the year we set out to prove whether or not mobile advertising can drive real world behaviour and not only did we do that, but we can now tell you how much it costs to drive an incremental customer in-store.

The quality of mobile targeting developed massively in 2014. The top providers now construct audiences using a plethora of rich data sources, whilst inventive start-ups push the industry to adopt new metrics and work harder to have a worthwhile impact on consumers. However, whilst previous years have seen great advancements in audience targeting on mobile, location data seemed like it was at a standstill.

2014 is the first time I think the market seemed to harness the potential. No longer was location data just about serving ads to people when they enter a location, instead we can now use location data to refine target audiences. In many instances observing repeat location behaviours -e.g. a consumer regularly seen doing the school run as an indication of being a parent- is much more robust than the inference commonly used in other digital targeting/audience modelling tools.

With options like this becoming available in the market, we chose to embark on the next stage. This was to prove that not only did mobile make sense in establishing a more accurate target audience but that the ads had an impact on the user and resulted in incremental visits in store. To do this, confidence in our methodology and the accuracy of data were the key pillars of our strategy. In this instance, the key to trustworthy results is tracking unique devices, and a robust enough control and exposed audience segment- the last thing you need when investing time and client budgets into something is to be let down by a control group of 14…

Clients of a decent size will always have other media noise around, and real-world behaviour is hugely impacted by factors like the weather, sometimes leaving regional testing with more variables than you started with- hence the importance of robust control and exposed groups. There have been tests on footfall before, although this has previously been done through vouchering, which in my opinion tends to tell you more about discounting than mobile driving real world behaviour. It’s also possible to marry Log-in data with your own CRM but this has only been good at looking for returning customers- what do we do about increasing market share?

To be confident our results were incremental we had to show that the same user who saw our ad subsequently went into store, whereas a user who didn’t see our ad was not motivated to go. This left a fair number of issues that needed addressing (a snapshot below):

-If we are tracking devices to store, what is a tight enough location to record them in? Is the common 250m fence sufficient anymore?

 -When tracking using in-store Wi-Fi, will we pick up users in the store next door? This shouldn’t be the case but it leaves a similar amount of doubt in mind to the 250m location data.

 -If we run on the tightest location data available (circa 12.5m), will there be enough ad calls for us to pick up a robust sample size?

Whilst the above is an indication of the issues we worked through, I think it perfectly shows the stellar advancements in mobile media over the last couple of years. That mobile media owners are equipped to discuss these questions and work with us on solutions is the big step forward.

Due to these developments we were able to deliver our client an ROI of x 2.7 on their mobile spend- and that’s excluding any digital sales from our activity.

For the strategy we delivered, the key measure of success was that we were able to show our client a cost per incremental customer in-store. This was all down to mobile location data, and for a client there is little more powerful in a meeting than having these numbers…and no doubt the questions it poses their traditional media counterparts.

Tracking consumers from ad to store is a very detailed real world digital strategy, and this level of granularity isn’t required for all. Finding the right target audiences, at the right time is surely a great start and a tick in the box for mobile location development. With a rich mix of consumer data touch-points now informing our audience segments, I would suggest that aren’t many more powerful than location data.

From my point of view, as a side project goes, mobile location data was a success- we picked up an award for our client on the way, and learnt more detail about the tech behind this than we ever thought necessary. We’re looking forward to the next steps and seeing what other solutions can be developed in the market- with so much bricks and mortar revenue still around, digital savvy clients are missing a trick if they let traditional media maintain an exclusive club.

If you’d like to talk mobile with Mark please contact him through the @MECUK Twitter handle or leave your questions in the comments below and we’ll drag him away from his mobile long enough to answer them.