Art Meets Science: Can Data and Creativity Coexist in Modern Marketing?
MMM and data that proves growth vs. common sense and optimising to remarkable outcomes over risk reduction
Mutinex co-founder Henry Innis and transformation consultant Tom Goodwin go head to head on “dirty” data, market mix modelling, channel planning and optimising to backside covering, versus relying on common sense, customer insight, context and risk taking to drive growth. They agree on one thing – but not much else.
Harnessing data at the start
Mutinex co-founder Henry Innis is rapidly scaling a media mix modelling platform that gives marketers a better read on how their marketing investments are driving incremental growth.
The alternative, for the last decade or two, has been marketing’s “obsession with shiny metrics” that he says are essentially meaningless. “When you correlate them to the P&L they do absolutely nothing, and that has caused huge issues with marketing’s credibility.”
Transformation consultant Tom Goodwin agrees that the “obsession with measuring absolutely everything we can” is a problem: “We’re just looking at dashboards, we’ve lost interest in our consumers’ lives, lost interest in the broader context.” He thinks there is “enormous value … in knowing what data to ignore” and not becoming reliant on data to justify decisions at the expense of common sense – and a degree of risk taking.
Innis says the key to harnessing data is using it at the start of a process, not the end.
“That is really important. I’ve seen a number of great creative campaigns that tested well, but they didn’t have enough media [investment] behind them to cut through – and that was the fundamental reason why they didn’t perform,” says Innis.
“It’s very helpful to have that sort of conversation, because it’s very easy to misdiagnose why things didn’t work and then react in a way that probably takes you down the wrong path. So use data to ground you at the start. Don’t use data to define where you go at the end.”
Data success has many fathers...
Goodwin’s not convinced that mining data can pinpoint why a brand is growing.
“My sense has always been that there are so many variables, and such poor quality of data collected about things that really matter, that while it’s possible to aggregate it. And while it’s possible to try and control for those variables, it’s pretty much impossible to get much meaningful data in terms of causality or inference or correlation from that stuff.”
A fabric softener, for example, will be sold into shelf space with some extra above-the-line spend. “They might have new packaging at the same time, a sponsorship of a sports team, then they might have a bit of a price promotion going on. The reality is, you’ll end up in this situation where there are perhaps 10 or 12 different variables. For instance, the economy, their competitor might do a price increase, a famous person they work with goes to prison,” says Goodwin.
“I’ve been in so many marketing meetings where 25 people will take credit for all of the success, but the moment something terrible happens it’s always the weather’s fault, or the macro economy.
“Most of the data we get is not particularly high resolution. Most of it is measured in fairly different ways. Most of it’s quite dirty. I think it’s possible to find lines of best fit, and it’s possible to infer things from it, But I don’t know if the quality of that data is so sufficient that we wouldn’t have been better off just drawing upon our own common sense.”
Innis is entirely “pro-common sense” but disagrees. Across the 65 large brands the firm works with in Australia, data may be non-standardised, but “most data isn’t that dirty”.
“I agree that causality is really hard to find in in any kind of correlation-type approach, but any kind of market mix model is an observational technique where we’re looking at one week, the next week, the next week, and at what happened in that week, what didn’t happen – things like that,” says Innis.
“If you get enough observations of something, you generally get a pretty good idea of what’s working and what isn’t . Could you do that 20 years ago with a regression model out of Excel? Of course not. But the world has moved on considerably.”
Juice versus squeeze
Speaking more broadly – i.e. not squarely about market mix modelling – Goodwin questions whether marketers might have lost sight of the woods for the trees.
“We can pay a lot of money for extra data, a lot of money in fees for platforms, a lot of money to technology companies to allow us to do new things.
“We now have all these meetings where we’re asked to present our data before we make any decisions. If you took all of that as a holistic approach and compared it with really good marketers 20 years ago, are they doing a better job? I’m not entirely sure they are, that the juice is worth the squeeze,” says Goodwin.
“People are afraid to make decisions. People feel like they need to create entire presentations full of data supported arguments to do things which are quite obviously common sense. And I think we’re entering an environment where we’ve lost quite a lot of intimacy with consumers. Instead, people spend all day looking at dashboards and making promises about what the return on ad spend will be.
“We’re in an environment where any discussion with data behind it looks more sophisticated, robust and progressive. I’m not too sure we should be confident that those arguments are more enlightened and more likely to lead to successful outcomes.”
Channel stupidity
But both Innis and Goodwin agree that talking about single media channels versus the plethora that actually contribute to growth is a fool’s errand.
“The thing I consistently see is it’s not one channel. Advertising is a cumulative effect of many things being active to then change a result over time,” says Innis.
Meanwhile, he believes the Mutinex data consistently shows the folly of overspending on lower funnel “performance” media to try to drive sales, “particularly when times get tougher”.
“In order to capture demand efficiently, you have to generate it efficiently,” says Innis, i.e. with brand investment. “What the MMM shows quite consistently is that if you just chase the bottom of the funnel it’s generally a pretty dangerous road to go down to generate long-term sustainable growth.”
Goodwin thinks music is the best analogy for ad campaigns, with media channels the various instruments that make up the whole.
“When you realise that marketing campaigns are like music, you realise that quite a lot of the conversations we have are quite stupid. It’s a bit like saying, ‘which is more important, the lyrics or the sound?’” he suggests.
“Someone says, ‘I think the lyrics are more important, so we should have twice as many lyrics,’ which is ridiculous.
“So the difference between streaming video and broadcast video [for example] is of no meaning whatsoever. A much better way to think about the modern media environment would not be to think of the media channel. It would be to think of the moment of the context of consumption. Are people sitting back? Are they paying attention? Are they open to new ideas? Does it feel premium? Are they trying to do something? Is it a place where they’re trying to do transactions? We need to think much more in terms of the broader context of the media, not the channel itself”.
Optimise this
Goodwin continues, saying: “My concern is that, generally speaking, the industry is much happier to make terrible music that appears to have followed all the best practice, appears to have been liked by all the focus groups, appears to stick to the same logic and the same rules of the past.
“For a lot of people it’s better to do something logically and in a way where they have something that’s not them to take the blame than it is for them to just make better music and to have taken a risk.”
That annoys him.
“We’re moving towards an environment where we’re trying to optimise for the reduction of risk, rather than optimise to do something remarkable.”
Have a question or comment about this episode? We'd love to hear from you. Please fill in the form.
