Stop trusting your digital data to improve strategic performance
We’ve been in the Big Data movement for some time, but now it is time to really dig deeper and be sure that what we’ve been measuring is actually reflecting the current and future customer.
It could be that reliance on existing data may be a limitation to opportunity. When the data of the past may have been captured with unconscious bias in place, or may be based on out-dated demographic profiling, it will surely be missing the modern customer. Many marketers and their agencies are not even questioning it. As the last step in the process, our creative work could even be doing brand damage. We need to rethink our reliance on past data across all data points in the marketing chain.
Many of the marketing and advertising industry are on autopilot when it comes to using data and looking at the analytics for their brand performance.
Those conventions that the media and ad agencies have relied on for a lifetime may be out-dated now, but nobody is stopping to check. The temptation with data is to rely on the ability for consistent comparisons through past metrics to justify the decisions made now. For many, it is hard to walk away from decades of precedence, when all we are doing as an industry is embedding a strategic misfire even deeper into our culture and society and limiting a brand’s potential. Much of this has self-interest at stake.
It is time to pull the green curtain up and question all our data.
How much of your data is based on the assumptions about family? Do you base decisions on the traditional model of Mum, Dad and 2.4 children? Well, think again. Only 30.3% of households are ‘couples with children’, and only 16.6% are ‘couples with young children. You could then say that over 83% are non-traditional families. That makes a focus on the traditional family a minority, and a strategic misfire for many brands.
Much of the shopper data that is embedded in the retail industry has the female Mum shopper as the primary target but interestingly many homes are filled with Grandparents taking care of children, single parents of all genders, and we have many that are actually an adult child taking care of an older parent even before we start thinking about same-sex families.
The current Australian Bureau of Statistics data includes all these types of families, but as an industry, we don’t. We also have not been able to effectively record the true number of non-binary or transgender population or same-sex marriages and families and the absence of data doesn’t make them non-existent. Only now will that data start to reflect a truer representation of our society now it is becoming more acceptable, and same-sex marriage legal. Reliance on past data is a big misfire right now.
Volvo is leading the way with this brilliant campaign in London that takes hold of this issue so positively and progressively:
There are many industry categories that traditionally focused on a masculine audience, and the organisations are filled with males in senior positions of authority still stuck in conventions and data of the past. We see these brands perpetuating the same bias through their communications in every possible way – from the strategic targeting, creative themes, tonality of the work, colour and font choices and the places we position the media.
Categories that were previously seen as the domain of the male such as automotive, finance, engineering, electronics, banking and so on, need to look past their existing data and really question if they are reaching their actual customer, and be ahead of their future customer.
General Motors Holden in Australia has recently released this campaign below featuring young men in an anthem style piece of work with all women featured as the inferior second fiddle, perpetuating an aggressive and dominant future generation of men lead by the example of an arrogant tween taking up the entire back seat while the girls are shoved to the dicky-seats in the boot. Someone must have taken existing data that gave the marketing department or brand the impression that their audience was male, and that women don’t have a say in the purchase of their vehicles.
According to Women Motorist in the US in 2000, Ford Motor Company determined that women have 95% veto power in automotive purchasing (data is a bit older here but it can’t have changed that much since). It was also revealed that $300 billion is spent annually by women on used car maintenance, repairs and service. If this data were available for Holden too, surely this campaign would never have been created? Not only that, it is so offensive and out of touch to women, it could be said that Holden is now in brand damage territory, with many women unlikely to touch it.
Putting that all aside, what data suggests that the age of the hero men fits with a teenage child in any case? What data shows us that a family 7-seater vehicle would be decided on by the male of the family, without any respect at all to the female in the family? It is hard to imagine there is any, or if so it is data misrepresented through a strategically misfiring creative execution.
The Justice Department’s National Institute of Corrections in the USA proved that artificial intelligence amplified existing racial bias by close to 200% through a system that was designed to predict the likelihood of criminals re-offending.
The system was biased towards Anglo-Saxons and allowed white criminals to be released, only to re-offend, and black criminals to be sent to prison and who were actually unlikely to re-offend. This is a great learning and a wake-up call for all our digital analytics to check if our data has any minority bias in place.
In larger organisations and enterprises, using AI for operational tasks or hiring protocols we’re getting better at preventing bias with proven algorithms available to help us eliminate issues like this. In marketing, however, it is still questionable that we’re even thinking about this.
Rethinking data in customer profiling
These are just a few of our industry challenges and the list can go on about the way we see people of different ages and life stages, differences between urban and rural populations, our customer’s sexuality preferences, and there is potential to tap into personality differences and thinking styles that is not yet even on the radar for many marketers and agencies.
It is also statistically significant that those with the highest net worth may well spend less than the lower socio-economic groups, that the average household income may be skewed, and focusing on salary levels to target a customer may be a misfire too. If we’re looking at data, we’re not looking at the right data in many instances.
One of the biggest ways we can rethink data is to stop measuring people by out-dated demographic profiling.
You will know when your media or agency partners are stuck on outmoded thinking when they say things like “That’s the way it has always been done”, or “That’s how we buy media”. Consider that the historic data set becomes invalid if we change the way we measure now, and the impact that has to media channels and partners with a vested interest in continuing the same old way of doing things will be a big motivator for sticking to the status quo.
There has to be a day we stop doing something that is wrong, and do what is right for the customer and the brand’s growth
More importantly, we need to stop harming our society by continuing to engrain broken concepts.
As marketers, we need to stop that thinking of the ‘chicken and the egg’ conundrum where we think we are there to simply follow conventional cultural cues. We’re lagging behind, and our lag is causing harm to brands and to our society.
Marketers and analysts need to ask better questions. If we have two customers defined by the typical demographics such as age, gender, occupation, salary level and socio-economic position sitting side by side we need to ask if both of them likely have the motivation to buy your product? If one will buy, and another won’t, then demographic profiling can’t be right. The waste in media spends, and the annoyance to the disinterested customer could even go as far as causing brand damage and certainly a missed opportunity for growth.
Rather than a horizontal slice of the consumer base using demographics in a hit and miss style, we can instead look to attitudinal segmentation.
We attract a very wide demographic profile that past analytics may not capture, and past media buying patterns may not align, but we reach an incredibly targeted and niche audience who are motivated to act and engage to purchase your brand. This new customer has a problem your brand can solve, and it doesn’t matter if they are male, female, 30 or 60, white or black, married or single, rural or urban, or how much they earn.
The new customer is highly motivated through an aligned set of beliefs, problems to solve, attitudes, and emotional drivers. We’re currently waiting until our audience has watched a strategically unsound piece of creative to make buying decisions and it is too late and too wasteful to get to that point. Think of that Holden campaign – losing a good portion of the female customer here, or Volvo gaining through their strategic decisions and reading of the data.
We need a new dataset that determines the motivating factors to purchase through customer research specific to your brand category. We need media profiling to be based on topic or interest groups instead of out-dated demographics. We need to remove all data that has been based on existing bias. We need to consider behavioural economics and to introduce qualifying steps in the buying journey, and even target 1 to 1 with dynamic content. At the very least, we need to get our creative more aligned with our actual customer without prejudice and bias in place, missing the mark strategically.
Brands need to take responsibility for the way they read the data and impact on society as a result of it and stop letting their agencies work on auto-pilot.
- Australian Bureau of Statistics 2016 (most current) – family population and socio-economic spending trends
- .id Population Experts – Australia profile https://profile.id.com.au/australia/households-with-children
- Women Motorist 2000
- AI Racial bias: https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing