I started my chat with Gurman Hundal, CEO of programmatic business MiQ, by asking him about the data challenges marketers are facing.
Hundal reframed my question slightly – “If you’re a marketer, data is your opportunity and your challenge,” he said.
On any other topic, this could have come across as the sort of hollow motivational sentiment one associates with LinkedIn. But Hundal’s point is so obvious it is often overlooked. The rhetoric around data includes frequent mention of the GDPR, organisational silos, the talent gap, legacy infrastructure, and the difficulties of measurement, but we shouldn’t forget there’s a reason marketers are wrangling these problems.
In Econsultancy’s 2019 Digital Trends survey, the most popular response amongst B2C marketers, when asked their ‘single most exciting opportunity’ in the year ahead, was ‘data-driven marketing which focuses on the individual’.
Furthermore, 2017 research from Econsultancy revealed that marketers at leading companies (those that exceeded their top business target) are more likely than their mainstream counterparts to say they make decisions about digital advertising based on data and analytics (71% vs. 51%). These leaders are also 35% more likely to have a documented data and analytics strategy.
Do not envy the marketer who leads an easy life.
Three challenges for marketers
Hundal is pretty clear on the three main facets of the opportunity/challenge for marketers – connecting data, discovering insights and driving business outcomes.
“Firstly, there is so much data now,” he says. “If you think about a client that has digitalised their business, they’ve got so much more data on their consumers who are interacting with their online platforms. Then there’s all the data that they can buy, there’s data on competitors, data on other potential consumers, all this is coming in different sizes and different volumes. So there’s loads of data to connect to get that full picture of their business, their market and their competitors. That’s a challenge.”
Then there’s analysis, the data science element, which is likely what most generalists think of when discussing data skills. Hundal says that “even if [clients] can connect their data, because of the volume, variety and different structures, they need data science expertise to be able to pull out insights and intelligence.”
“The third thing they need to do is to make sure that intelligence is actionable, that they are using it in all their communications channels as close to real-time as possible. That’s an engineering problem too,” Hundal concludes.
Data scientists don’t grow on trees
Data skills are in high demand across every industry. A Burning Glass analysis of UK online jobs listings found that in the five-and-a-half years to June 2018 there was a sharp rise in UK job-listings for ‘Data Scientists and Advanced Analysts’ (+ 231%) driven predominately by increased numbers of vacancies for Data Scientists (1,287%) and Data Engineers (452%).
The same report notes that “with major industry players hiring many of the most experienced data scientists and AI researchers, media reports have suggested that the natural flow of researchers from academia to industry may be reaching unsustainable levels.”
These ‘major industry players’ are also name-checked by Hundal when discussing the talent gap. “That sort of talent tends to want to work for [tech] platforms…” he says.
Interestingly, he adds that hiring data engineers and scientists is “culturally a big challenge” for marketing departments. “They need a different kind of skillset” and “the opportunity with data has got to be realised in the marketing department.” But that’s no mean feat, even if a marketing team already has some programmatic or analytics capability in-house.
This is the context for MiQ’s work and is why, within 10 months of the business’ launch, a business entity in Bangalore was set up to get house this requisite expertise in engineering and data science. Hundal tells me 50% of the workforce is based there and that they are the “nucleus of MiQ.”
Using this expertise, MiQ runs effective programmatic display ad campaigns for clients, and it does so by tackling the aforementioned areas of connecting data, applying data science and then utilising the outputs. Hundal describes it as the “concept of marketing intelligence” and says “It’s not just targeting, there’s value in giving companies that consultancy elsewhere, in feeding back some learnings.”
When I asked Hundal about the maturity of this data science, he agrees “there is still a gap between [martech] pitching [AI and machine learning] and what their actual capabilities are.” Indeed, an eyebrow-raising stat came from a study by London venture capital firm
MMC early in 2019 – 40% of European startups that are classified as AI companies don’t actually use AI in a way that is “material” to their businesses.
However, Hundal is pretty pragmatic and outlines the basics of predictive analytics as an example. “The reality of what you’re trying to do is predict what your consumer is going to do,” he said. “In digital programmatic media it’s a really important technique to find new customers by forming an assumption to decide how you will target your campaign.”
He talks through one of his favourite hypothetical examples – “If we had an assumption that people buy more US Open tickets when it’s 28 degrees outside and there’s positive sentiment on Twitter about Roger Federer, for example, for us to come up with that assumption, we would have to analyse all sales data, all the characteristics of users, all of the variation in tweets, the weather conditions in every area. …When you’ve got big data and big variety, that’s where AI comes in – it’s the data science technique, effectively, to analyse some of that data and pull out predictions. Then make sure we buy that way.”
Agencies are beefing up their own data capabilities
Over 90% of MiQ’s revenues is from agencies, roughly split 50-50 between the six big holding groups and independents.
According to Hundal, data is “at the front and centre” of “strategic moves” being made by media agencies right now. “They are trying to be strategically valuable to their clients by being right in the middle and controlling their data assets,” he says. He points to the acquisitions of data companies Acxiom and Epsilon (by IPG and Publicis Groupe respectively) as prime examples.
“They’re all trying to, not just buy data, but buy in that capability to offer that source of analytics for clients, where if the marketing department don’t want to build that, agencies can offer that with media buying.”
When I question Hundal about whether marketers can be forgiven for being wary of media agencies potentially having this greater level of control over data, he is clear “it’s all about being agnostic.” He cites WPP as an agency group doing a good job, and comments that “they’re not trying to create an access-led point of difference, they’re trying to be the middleware and connect whatever the data assets are.”
He continues: “If an agency tries to say a data asset of theirs is the single point of truth for a brand that’s a conflict. So they still have to have a service led approach. Similarly, if an agency’s trying to prove their activation channel is the only one that works or that their data there is a conflict. Then they’re a supplier and not an agency. They need to be agnostic and license in whatever asset will be of value to the client.”
The next generation of programmatic innovation?
Whilst no advertising medium is particularly loved by consumers, it’s no secret that programmatic display has had its detractors when it comes to creative. When I put this to
Hundal he is both realistic and sanguine, believing that display may have its weaknesses but that programmatic is only just getting started.
“In the last 10 years, it’s been mostly restricted to a website on a desktop/mobile/tablet. …And yes, TV and outdoor, the engagement and creative always captures more attention – a video ad on a publisher website has content all around it and it’s often a small section of the page. But, looking at the future, I think the next generation of programmatic innovation is how those creative executions in radio, TV and even outdoor are going to be bought programmatically because consumers will consume those via the internet.”
Hundal highlights the decline of broadcast media amongst younger age groups, and there are certainly studies which reveal the penetration of digital radio and over-the-top and connected TV services. Ofcom’s Media Nation report showed that among 16 to 34-year-olds, total daily viewing time in 2017 was 4 hours 48 minutes, of which less than half (two hours 11 minutes or 46%) was to broadcast content, while just under an hour per day was spent watching content on YouTube.
“If you can overlay some of that clever programmatic targeting [on this creative], why wouldn’t you?” asks Hundal. “Everything now is about incrementality – ‘What was the biggest contributing factor that caused sales to increase?’
This is the dynamic that led MiQ to develop a product in the US called Cast. The company website states that IP integration enables a nearly three-quarters success rate in matching TVs to an online device. This connection between TV viewing behaviour and digital data from another device can then inform activation across media buying and other channels. It’s not hard to envisage such integration becoming a quotidian part of modern marketing.
The marketing team of the future
The good news for agencies is, as Hundal puts it, the marketing team of the future will still need “strong creative people” and “people who are very strategic and understand business needs.”
“But when it comes to activation and execution I think you will see a lot more data engineers, data scientists and people who can activate on mediums which are all digitally enabled,” he adds. “In 10 years you might argue that every medium is digitally enabled.”
Hundal concludes the future is bright for agencies because those that are “making the right moves to be data competent and offer value to their clients” are giving CMOs “exactly what they need.”
Strategy, creative and skill – in a data-led infrastructure. Looks simple on paper.