For years marketers have complained that they aren’t seeing the returns from social media. Another year and another Duke University CMO survey finds that barely one in 10 marketers feel that they are accurately measuring the returns from social.
They haven’t adapted.
Allow me to explain.
While it’s possible to use traditional FMCG-style brand tracking models, as Nielsen does in partnership with Facebook, social budgets are usually significantly smaller, and so struggle to justify the research costs.
Other marketing channels, from email to direct mail, have tended to have a clear and direct relationship between marketing activity to conversion. So while you may open an email with an offer from your favourite fashion retailer on your mobile, you will use your laptop to buy that dress, clicking through again from that email. Similarly direct mail will have a unique phone number reserved for people who have received that mailer.
But in social it’s very different. Around 80 per cent of all social media consumption is on mobile devices. And it doesn’t persist, because almost all social is stream based. So when people decide to buy something they have a choice.
They occasionally click straight through from a piece of content to a website and buy something. But it’s fiddly to buy things on mobile – with usually at least seven form fields to fill in – so only a fairly small proportion of people do so.
The much more common alternative is that people simply switch channel. So having seen a good offer on that dress on Facebook, you go to your laptop, Google the retailer and arrive at their site as a direct visit.
In the 2015 ‘State of Mobile Commerce’ report from Criteo, it was suggested that at least 40 per cent of ecommerce converts on a different device from where it started. 40 per cent!
Perhaps this is intuitively obvious given the ease of larger screened devices for filling in forms and undertaking complex actions, however the consequence is that huge amounts of commerce is misattributed to direct website traffic, when it actually originates from social.
This means that last click measurement of social’s impact is fundamentally flawed – because it measures a behaviour that most normal customers haven’t adopted.
How big is this? It differs a lot by category but working with client data at OgilvyOne we’ve seen plenty of examples where 10 times more people would sooner switch device than convert directly on social. A recent Radium One estimate stated that over 70 per cent of UK social traffic is unmeasured this way.
This measurement difference is transformational for social.
What are the solutions?
Broadly there are two.
Firstly there’s cross-device analytics – in other words joining up these journeys. Much easier to do in the travel industry for example, with everything from mobile booking passes to trackable login details (via desktop/laptop and mobile), but hard in practice for many other categories. Facebook and Twitter’s conversion tracking pixels also make this easier – again if you’re a brand where a lot of your sales are online.
Most obvious areas for marketing analytics not used by most marketers. Source: CMO Survey 2015 pic.twitter.com/pnAk4vkkRT
— Rob Blackie (@robblackie_oo) March 30, 2016
Secondly, you could go down the route of excluding a portion of your audience from your social marketing and measure the difference in purchasing rates between people exposed to your social activity and those who are not. Again, Facebook and Twitter are useful here as their Custom Audience products now make this exercise relatively straightforward. If you have a large customer database it’s quite easy to apply this to offline purchasing as well as online.
If it’s that simple, why isn’t everyone doing it?
Partly this is inertia. Organisations struggle to change, and a substantial proportion haven’t even adapted to mobile yet. Unbelievable but true.
What’s worse is that even if people want to change it’s too difficult. The amount of times I’ve heard the answer to "Do you have conversion tracking installed?" as "The legal department won’t let us" is just depressing. It’s not a great explanation, but changing anything takes time, and most organisations are struggling to work out a reasonable position on privacy (and fair enough, it’s complex).
However, I would argue that this is a failure of the social industry as a whole – far too often social has not demanded to be measured on results. And, as an industry, we should collectively fight to measure social properly.
The tools and the methodology are available.
We just have to grow a pair and use them.
Or else, what’s the point?
Rob Blackie is director of social at OgilvyOne. He tweets @robblackie_oo