Joining the data dots to predict the future: Meet Ometria’s Ivan Mazour
With nauseating velocity I’m elevated 40 floors up the Heron Tower, my ears popping with altitude, to meet Ivan Mazour at the Duck and Waffle.
We’re here to talk about how Ometria, his seventh company by his early thirties, can predict the future.
Often referred to pejoratively as “the son of a Russian billionaire”, Mazour is very much his own man; loquacious yet shrewd. “I’m a big fan of talking – and talking quickly,” he says with pace, as I tuck into seared octopus, and one of his unwanted waffles.
Ometria is a customer insight and cross-channel marketing platform with a twist. It analyses first party data: past store visits, online purchases, interactions and behavioural insights, to plot a more advanced digital map of an individual. When run through machine learning algorithms it will attempt to predict that customer’s next purchase, and target in a way that is highly-personalised, non-intrusive, and across multiple platforms. Simple, right?
Carpet bomb
Mazour insists that the old-school carpet bombing by marketers – sending mass emails in vain hope of hitting a few relevant consumers – is more burden than boon. “There’s a huge transition that’s happened over maybe the last four or five years, of consumer expectations; of what they will and won’t put up with when it comes to the way that brands and retailers communicate with them.”
He says the consumer needs to be made to “feel special”, and to do so, “fundamentally, you have to first understand them. That’s true in human, direct relationships. And it’s true in a retail and business environment as well.”
It’s a confluence of crashing loyalty and heightened expectations that Ometria is addressing. “If the expectation to be treated as someone particularly special is so high, and the ability to go somewhere else is also so high, the retailers, they’re being squeezed harder and harder,” says Mazour. “In a perfect world, you would have in place an understanding of what it takes to keep your customers loyal, and you would have in place a way of communicating with each of those customers to remind them that: ‘I remember, I care about you, I want you to stay loyal, and I will do everything I can to do it.’”
First party data offers far greater accuracy across platforms (Source: Getty)
Retargeting
Ometria’s solution challenges a bugbear of marketers and consumers alike: the irritating retargeting of unrelated adverts. “It’s a huge problem in advertising today,” says Mazour, clearly about to embark on a diatribe. “I’m wearing these shoes right – I keep talking about this and I should really stop – I bought these shoes on a random Saturday, but I’m still being advertised them six months later. And they know I’ve bought them – of course they know, they have that data.”
The difference between Ometria and its competitors is its use of first, as opposed to third, party data. It knows its clients’ customers by name, which adds to the hyper-personalisation and to making the consumer “feel special” that is so central to Ometria’s mantra.
Using data already held within businesses, says Mazour, can unlock new insights.
“Every single marketing or advertising technology provider over the last couple of decades has focused on the third party data set. So basically they ask: ‘can we connect together some cookies over a whole bunch of sites?’ And maybe I’m looking at blue jumpers here, and holidays in Cuba there. Link it all together and try and build a profile.”
But this method is facile, he says. “The problem is that it’s never identified down to an individual – it’s the whole point of third party data, staying away from all of the data privacy issues that could potentially arise with that. And the problem with that is, if I’m doing it across different devices, it’s practically impossible to link those together. If I’m doing something in a physical store, it’s impossible to link it together. There’s all these touch points that can never be connected in a third party world.”
Like a data dot-to-dot, Ometria connects those points to glean predictive insights about consumption patterns – essentially, to predict the future. “Just having a data set doesn’t help the retailer to understand what I [a consumer] want next,” says Mazour. “So we run a whole bunch of machine learning algorithms over all those data sets to work out, for each customer, what they are interested in buying, when they are going to be interested in buying it, and what is the best way of presenting it, so they make that purchase.”
Once it has figured out what to send, the final piece of the Ometria jigsaw is omnichannel delivery. “Marketing technology in retail for the last 10 years has basically focused on sending emails. But the consumer of today expects a unified journey across all of the channels, expects it to be super-personalised. Every message to them has to be different to someone else, and delivered at the right time.”
But how accurate are the predictions? “I feel, when you look at our consumers, we have over 20m customers profiled, and at that sort of scale, the overall effect is extremely accurate – you might get people wrong by a week here or a week there – but the messages are always relevant, each consumer’s experience is already better than it would have been before.”
Elliott Haworth is business features writer at City A.M.