Image by Midjourney, prompt by dominikwieschermann

Will personalization fail (by 2025)?

In 2019 Gartner predicted that personalization will fail by the end of 2025 (1). Here is why I think they are still right, and how we can fix it.

Personalization promises both increased value to the customer and impact to the business. But it all hinges on the first part: value to the customer — and this is where personalization easily fails.

a. Where personalization fails

What is personalization and how do we get there?

Commonly organizations have been sold on the narrative that behavioral data like click-throughs, opens, content tagging, conversion etc. are all valid proxies helping us understand and include our customers.

They aren’t.

The only thing this data: web channel performance data, helps us do is optimize the design of our own channels.

Behavioral data is a window into the effectiveness of our own designs, NOT our customers situations, emotions, culture, relationships or needs.

Showing the difference between thick (subjective, motivational) and thin (behaviors) data. (3)

We can of course make questionable assumptions (2) about our customers based on how they move around when they are with us, but we should be aware that:

• Platforms are most often designed through the lens of our own organization and products. So they are natural to us, but artificial to our customers.
• Understanding what people do or why they are doing it are two very different types of insights (3).
• We don’t know what they are doing everywhere else. And likely the time they spend with us is not significant or representable and might produce a bias on what we are learning.

At the heart of why personalization fails is a story we have been telling ourselves for years: that our own channels and our customers are the same thing. Creating the questionable assumption that our own products or websites are a reliable reflections of what people need.

b. A new personalization mindset

I would argue that most personalization-efforts today are not focused on the customer at all, rather they are product or channel-myopic. The focus is on how we build to individualize our content or channels rather than improve how we serve our customers.

Assumption: A customer doesn’t care if she is getting the same experience as one or a thousand other people. What she cares about is what she needs and how she gets it.

Personalization should mean that we are better at understanding our diverse customers and able to tailor how we show up to better serve each person the way that is best for for them — leading to an improved outcome for the organization.

This means understanding, appreciating and designing for the complexity of each individual human being.

It means measuring if our customers are getting their job (4) done better, compared to treating them all as the same. And if it’s worth the business investment.

Therefore personalization cannot start with our marketing automation systems, our advanced analytics or modular content. It has to start with better understanding our customers.

How: we need to be clear on what business impact we want from personalization then find the important unmet customer needs where personalizing the experience would increase the value to the customer and the added investment pay off back to the organization.

Start with the business impact you want, find the customer behaviors that drive this impact and what needs and outcomes motivate it. Where would personalization make a difference? Illustration by the author

c. How do we fix personalization?

Simply put: the organization needs to realign itself with its customers’ needs and creatively pursue data to that end.

My assumption is that the reason we are looking at personalization today through a product and channel-myopic lens is because we are using the data most easily available to us.

We have to get new data, and we have to be creative about it. Most data is not directly in front of us.

e.g. some years ago Aviva, a British insurer discovered that people’s unconventional data such as online behavior and spending habits was as effective at identifying future health risks as a medical examination including blood and urine tests (5).

We can design for the data we do not have. Online experiences are awesome laboratories to the learn deeper / thicker (why) insights about our customers .. If we design for it (6)(7)(8).

Lastly we overcomplicate what we need to know. Nobody shows up as a full human being in an experience, people show up in the situations they are in. we only need to know why they are with us right now, and what they are trying to get to next. (9)

The problem with personalization is not that it won’t work. In fact it has always worked, but we scaled ourselves away from understanding it. We forgot the human component of our experiences and offerings and started focusing only on our own products, channels and content (which makes it very simple, but wrong).

d. Three simple questions

In order to check personalization ask yourself these three simple questions:

1. When successful, what is the customer and business impact?
2. What will be different (compared to a generalized experience)?
3. Which decision and measures do we need to be successful, and what data / insights do we need to support these?

Sources:

(1). Gartner Predicts 80% of Marketers Will Abandon Personalization Efforts by 2025, https://www.gartner.com/en/newsroom/press-releases/2019-12-02-gartner-predicts-80--of-marketers-will-abandon-person

(2). “Models are opinions reflected in mathematics” — Cathy O’Neil, https://everythingnewisdangerous.medium.com/models-are-opinions-reflected-in-mathematics-oneil-a93ad607f893

(3). The customer data gap, https://everythingnewisdangerous.medium.com/the-customer-data-gap-520cdf695d68

(4). Know your customers ‘Jobs to be done’, Clayton M. Christensen, Taddy Hall, Karen Dillon, and David S. Duncan, https://hbr.org/2016/09/know-your-customers-jobs-to-be-done

(5). Risk and Reward, The Economist, https://www.economist.com/finance-and-economics/2015/03/12/risk-and-reward

(6). Design for better data, https://bootcamp.uxdesign.cc/design-for-better-data-3dbc7fef28c5

(7). If we design for it, https://everythingnewisdangerous.medium.com/if-we-design-for-it-34e37d067746

(8) Designing for learning, https://everythingnewisdangerous.medium.com/design-for-better-data-6d2f780028d

(9). Customers don’t buy products, situations and needs do, https://bootcamp.uxdesign.cc/customers-dont-buy-products-situations-and-needs-do-520b5491f8d1