Maybe You Could Be Doing It Better

Your customers may be more diverse than your marketing.

How well do you know them? Do you have your target segments and do you build marketing campaigns to reflect what you know about them?

Many brands divide their customers into groups based on common characteristics so they can market to each group effectively and appropriately based on those pre-determined characteristics.

But how do you know if those are the most accurate groups to engage? If your sole method of understanding your customers is through demographic segmentations, then at best your understanding is limited and at worst it is incomplete or even misleading.

We vote for letting the data itself reveal the customer personas that inherently exist.

First, let us add some context to why this is important.

We are currently living in what we call a “post demographic” world. What we mean is that consumers have changed. The way we all interact with brands has evolved considerably in recent years. Clean lines between consumers based on gender, age, income, or ethnicity are not as useful as they once may have been.

A female working in finance with more disposable income than her male counterpart working in education very well may have more in common with their values and lifestyle than just demographics would suggest.

People across the world continue to construct and re-construct their own identities and rebel against top down driven norms given to them by advertisements of old.

This world, where commonality is not necessarily based on demographics or other control variables, has massive ramifications for the accuracy of the future of consumer segmentation and by extension, marketing and advertising.

Supposedly you conducted research to understand the likelihood to purchase a new consumer household good. You collected data from 1,000 respondents: sex, age, geography, and sensitivity to things like price, brand image, product quality.

A traditional consumer segmentation may have revealed that females are more likely to purchase than males. Or that when you cut the data according to generations, you found out that Gen Z cares more about brand image than Gen X.

But is that the only way that these consumers are similar?! How old they are?

Let’s circle back to our original suggestion and allow the data to do the talking. We do this by using a type of machine learning algorithm called unsupervised learning. This allows us to segment the data according to how the data is behaving and cluster consumers in the most homogenous and efficient groups possible.

The results may show something like the below, a three-dimensional cluster analysis from a product R&D study where the consumers in each group are similar to each other based off of how important quality, image, and price were, and not based on a pre-determined demographic split. Each persona tells a different story.

The first cluster scored low on sensitivity to price and higher on both product quality and brand image. The second cluster scored much higher on price sensitivity, low on caring about product quality and lower on brand image, and the third cluster scored medium about price, and the highest on caring about both product quality and brand image.

The interesting part? Each cluster was made up of a mix of demographic variables!

Each of these unique personas can be messaged to slightly differently. You need to understand how your consumers view the world so you can meet them on their terms.

The purpose of learning is growth.

• Consumers and their behaviors are constantly changing and evolving. Conduct segmentation frequently to to make sure your marketing is up to date and on target;
• Adapt your marketing strategies to the motivations and desires of your newly discovered segments. By also collecting data about on-line and social media habits you can refine your strategies even more;
• Demographics don’t tell us the whole story. You need to collect data related to who your consumers are, what their values are, what they care about, and what drives them and motivates them;
• Complement traditional consumer segmentation with machine learning. Every statistical method comes with its own assumptions. Even “no assumption” is an assumption;
• Be open to learn about results that confirm your hypothesis and your strategy, and other findings that may surprise you. Staying relevant to your customers to build long-term brand equity and loyalty. Learn, adapt, grow!