Prophet CEO Jordan Taylor-Bartels and COO and VP of marketing science Hamish Mogan urge marketers to stop thinking of media spend through a linear approach. All kinds of things, from the geopolitical climate to the movies we watch, can influence a consumers’ purchasing behaviour. So why are so many marketers still lost when it comes to predictive marketing analysis? Taylor-Bartels offered up solutions, from digital clones to flight simulator-like approaches, on his panel at the Google Cloud Summit, held at the ICC in Sydney.
Prophet launched to market just a year ago, and the venture has gone better than expected, according to CEO Taylor-Bartels, who chatted with B&T after the panel.
With 30 clients already up their sleeve, from BYD to Channel 7 and life insurance company TAL, and others in the pipeline that Taylor-Bartels can’t share just yet, the company has already made moves in the industry.
Prophet raised $5 million in funding last year and is being backed by heavy hitters, including Antony Catalano, executive chair of View Media Group and former Domain CEO; Seek co-founder Matt Rockman; and Bastion CEO Cheuk Chiang.
“The question we’re really asking at Prophet is why the odds feel so stacked against us in marketing. One reason is that the market landscape changes so rapidly. You judge good decisions by how quickly you can make them, but also how well-informed those decisions are,” Taylor-Bartels said.
“Particularly in marketing, we’ve got publishers and agencies pulling clients from side to side, saying, ‘hey, spend more on this channel’. These decisions are being designed by what we call media mix modelling. This model has been the guiding light to how marketers and organisations can begin to understand more about how their entire channel mix is having an effect.
“That’s all well and good, but media mix modelling was designed and built almost 20 years ago. The world was a very different place, in terms of data, but also in terms of geopolitics, in terms of everything we’re seeing today,” he said.
Harnessing data to inform decisions
“A lot of CMOs don’t have confidence in their strategies; they don’t understand which way they should go, let alone understand their own personal instinct,” Taylor-Bartels said.
“Sixty-eight per cent of board directors admit generally that they don’t know that what they’re saying is actually translating into some kind of effect. This isn’t because any of these people are bad at their jobs. It’s because they’re operating in an incomplete ecosystem with incomplete data, and they don’t quite understand how it all works.”
Delayed insights about a huge mix of data can create a lot of confusion, he explained. Relying on unorganised and delayed data insights to inform decisions can end up feeling more like solving an algebraic equation and less like actually connecting with consumers.
Taylor-Bartels said the key to ensuring more CMOs don’t get sacked is to break knowledge barriers.
“Real-time data is now the focus for MMM specialists,” he said. “Linear approaches don’t work for most businesses. MMM models use data captured from November to make decisions for July, and by then, too much has already changed”.
Taylor-Bartels honed in on the use of AI in Prophet’s model. While AI can’t be used to predict things, it can summarise and present what humans have already done. But this technology can be harnessed to organise datasets in such a way that the humans behind Prophet will be able to make highly accurate media predictions.
One way this has been done at Prophet is through digital clones, which allow marketers to run alternate yet plausible simulations of the future.
“We want to enable you to make better decisions tomorrow, rather than just retrospectively saying you did a good job yesterday,” Taylor-Bartels said.
The digital clone model built by Prophet harnesses a vast array of data that can adapt to the market as it’s moving. It is on-demand and can be refreshed by marketers to test, predict and optimise before they spend a cent.
“This philosophy came from our insight that pilots don’t learn how to fly by crashing planes into the ground each time. There’s no reason why organisational decision-making should be any different,” he said.
Mogan added: “Our digital clone takes in a number of different buckets of data: media data, offline data, paid search. But we also need to understand what contributes to that acquisition pathway. What competitors are doing, where our client’s brand is positioned in market, and anything else that can help us give context around the commercial outcomes for that business.
“The last piece of the puzzle, which is probably one of the most important, is the macroeconomic environment in which the data lives, which we bring to the party through 86,000 different variables, which we call the prophet library. And this covers a whole bunch of different things, such as interest rates, weather, equity, currency, demographics, you name it, a huge swathe of data that helps us describe that macroeconomic environment.
“From there, we build our digital clone using Prophet AI, and we spend a lot of time perfecting this craft. Our team spends an inordinate amount of time making sure that we combine and make this digital clone effective and reduce the error to such a point that it can be commercially relied upon.”
Enron 2.0
Taylor-Bartels emphasised the huge responsibility the Prophet has in advising companies how to spend their money.
“Behind these decisions is a reputation; if we’re telling someone they need to spend money in one area, and we get that wrong, that’s potentially a brand on the line. It’s really important that we’ve got the data and mathematics right to advise with extremely high accuracy,” he said.
He also wants to ensure the company remains independent. With huge players influencing the market, notably, Google, Taylor-Bartels emphasised that it’s imperative for agencies not to be influenced by them.
“Google is kind of putting its lobbyists into different points to make sure they can influence decision-making on measurement,” he said. “It has this big search P&L that’s at risk, right?
“I think it’s really important for agencies to make exclusive decisions, to say, ‘no, we need to work with a third-party provider that’s not going to be influenced by how these decisions are made’. This is kind of Enron 2.0; all it takes is one agency being manipulated by Google to give you the wrong decision-making, and it could tank a company. It may not tank the company, but it may mean five people get made redundant.”
Marketers need to be on their A-game when it comes to MMM. With so many advancements to the model currently being fine-tuned, companies like Prophet are well positioned to empower brands to take charge of their media mix decisions.
