The Upside of Personalization: Short-Term and Long-Term

← Back to all posts

In the previous pieces in this series (Part 1 here, Part 2 here, Part 3a here and Part 3b here), we discussed the typical ways brands can realize growth, came up with a function that attempts to describe the opportunity for personalization, and derived some specific equations that could provide an estimate of the upside. (They were very complex and I won't repeat them here, but worth going back to Part 3a and Part 3b to review.) Essentially:

Now I'm not claiming that this is entirely ground-breaking thinking, and that struck me as I was working hard on deriving the example formulas that could describe the impact of AI-driven personalization. As mentioned previously, substitute "promotions" or "segmentation" for "personalization" and the model may be essentially the same, and others have previously done the work of precisely defining the components.

Going back to the original entries in the series, it still seems hard to accept the McKinsey & Company or Boston Consulting Group (BCG) stats at face value. They don't suggest a timeframe over which these types of gains might be realized, and these "rules of thumb" about what might be achievable can't be applied against individual categories without some adjustment: they don't take into account category growth or a consumer's willingness to switch brands or the nuances impacting brand choice for less frequently purchased, larger items, among other shortcomings.

The big question that remains is the following:

Is AI-driven personalization enough of a game-changer that it can actually create demand where it otherwise would not exist, or is it just another tool that can drive short-term results for first-movers but not sustainable long-term advantage?

This is not a new question that has arisen due to the unique opportunity afforded by personalization, and in fact, there are some terms for what seems to be going on here. When a company is an early adopter of a new technology or trend and then the rest of the market follows suit, this can be referred to as "crossing the chasm" — a term originated back in 1991 by Geoffrey Moore (Amazon), itself an adaptation of Everett Rogers' "diffusion of innovations" theory from 1962, where the "chasm" represents the gap between early adopters and the larger early majority market segment that needs to be convinced to adopt the innovation.

This may also be known variously as the "technological equilibrium dilemma" or "innovation imitation trap" or "innovation inevitability paradox": once a technological breakthrough is introduced in a competitive space, others are compelled to adopt it to remain competitive, leading to an eventual equilibrium where no one maintains a lasting advantage.

The speed at which equilibrium is achieved depends on the difficulty in implementing the technological breakthrough — in the case of AI-driven personalization, there is no seeming technological barrier to adoption, only resource allocation limitations, and most importantly in the context of marketing, the availability of first-party data.

A general example of the "crossing the chasm" phenomenon that is often cited is the shift in basketball strategy towards three-point shots: teams historically relied on mid-range shots and post play until the Golden State Warriors deployed three-point shooting as their primary strategy, which initially gave them an edge. Ultimately other teams adopted the same strategy and the advantage disappeared, and three-point shooting became a necessity.

A business example might be the adoption of same-day delivery by online retailers. Amazon led the way, and Walmart, Target and others ultimately had to follow, either by building their own capability or by partnering with last-mile services like InstaCart or DoorDash. Online ordering combined with fast shipping is no longer a differentiator, but has become a consumer expectation, eliminating Amazon's advantage on that dimension.

In essence:

1. Early Adopter Gains Advantage
2. Competitors Are Forced to Follow
3. Equilibrium is Reached

Doesn't this seem to capture exactly what is going on with personalization? There is a short-term race to implement a solution, in many cases without even thinking about competitive response, but eventually everyone is going to try to catch up.

What this suggests is that there is no one blanket metric you can apply to estimate the potential upside; it is going to be category specific, and will be more impactful for higher-value, more involved purchases. More to the point, outside of larger, less frequent purchases, first-mover advantage will likely provide only a temporary gain, until competitors respond and equilibrium is reached, with certain exceptions where one brand quickly achieves dominance and is able to build BCG's "competitive moat". This will most often happen where a brand with dominant market share is able to leverage their advantage in first-party data to help refine their personalization, achieving gains that secondary brands will never be able to match.

Let's consider a couple of category examples to illustrate the challenge.

• Casual Dining. Recently, this has been a slow-growth/no-growth category, and only Chili's seems to be delivering outstanding results. That means any gains Chili's is realizing must be coming at the expense of competitors. (Industry analyses in this particular case cite promotional strategy as the primary driver of Chili's gains.) The upside of personalization in this category will always be limited because it will be subordinate to key decision factors like family budget, need for variety, convenience, etc., and the ability to personalize will be dependent on the availability of sufficient data to drive personalization in the first place. If you don't have a mechanism to capture first-party data, you're going to miss out entirely. And even with that in place, it will always be challenging to overcome category dynamics.
• Automobiles. The opportunity to personalize the sales approach is tremendous in this type of category, but limitations exist because of a lack of unique data due to infrequent purchase cycles and changing household dynamics over time. This means using third-party data may provide a short-term advantage, but that will dissipate quickly since the same data is readily available to all competitors. Many consumers may start with a loose set of requirements, and more likely have a small consideration set of acceptable brands already in mind based on observation, word of mouth, status, etc. Personalization would have to overcome these biases for a brand not in the initial consideration set to have a chance.

This simple exploratory exercise certainly caused me to really think about not only the potential upside, but the importance of putting it into temporal context, which has been omitted from most if not all of the articles and books I've reviewed. This may be a contrarian view given the buzz around AI-driven personalization in the marketing ecosystem, but here goes:

• Short-Term: There is definitely some first-mover advantage opportunity and leaders in personalization can expect to see positive impact on spend per customer and market share, with magnitude dependent on positioning and category dynamics. Impact may not be as dramatic as McKinsey or BCG suggest, but something.
• Long Term: Short-term advantages will tend to dissipate in most categories as competitors achieve parity in terms of personalization. Businesses with direct consumer relationships and dominant market share, especially those that serve multiple frequently-purchased categories, may be able to sustain some level of advantage.

So to answer the original question: what is the upside of AI-driven personalization? In the short-term, it is a function of category growth, ability to expand consumption, and the likelihood that consumers will stray from their preferred brand for a given purchase occasion given personalized messaging and incentives. So mostly not that dramatic. Over the long term, it is a function primarily of market share and category growth, with the largest spoils going to those brands that can capture new category entrants.

This means a short-term race to build first-party data collection mechanisms, and to solidify your approach to personalization. If you haven't already, get going.

Then it will be on to the next shiny new thing that's going to make a difference. And I'm going to go watch Minority Report for inspiration.

View original on LinkedIn