Having worked across a range of industries, I feel lucky to have been able to observe the effects of personalisation across a wide range of channels, moments and marketing life-stages. 

From the e-mails that airline loyalty programmes send you about where you flew in the last year, to the curated offers that supermarkets send you based on your historical shopping behaviours, to chatbot conversation flows that are designed based on predictive recommendations on what you might be interested to learn more about. 

I feel comfortable saying that categorically, personalisation can unlock significant, scalable and repeatable business benefits. I’ve seen the results enough to believe in it.

However, the biggest challenge that prevents businesses from realising those benefits is generic strategy when it comes to personalisation.

Too many businesses seem to have applied near-identical “Crawl, Walk, Run” frameworks that first start with ‘broadcast’, move to ‘segment-based’, then start building ‘journey-based’ personalisation… whilst eventually ending up at true 1:1, true real-time, contextual, predictive… <INSERT BUZZWORD>, AI-driven mega personalisation.

Lured by the potential benefits of personalisation, we’ve all been working towards a utopia that’s been sold to us.

But while we’re all working towards this hyper-personalised end-game, we’ve neglected the importance to stop and personalise our businesses approach to personalisation.

Here’s my provocation: Does every business need true 1:1, true real-time, contextual, predictive AI-driven mega personalisation?

I would argue no, and to explain why, I want to introduce what I’m calling “personalisation tolerance”.

Personalisation tolerance is the idea that organisations need to define their own thresholds at which personalisation efforts become counter-productive to their ROI. 

This is also an idea that organisations need to define their own scales of maturity, relative to reasonable aspirations for their organisation, not generic scales for their industries or even industry as a whole. 

So, what is a ‘personalisation tolerance’?

A personalisation tolerance is a limiting factor that sets a threshold at which personalisation efforts become counter-productive.  Organisations benefit from initial efforts, but beyond a certain point, they yield diminishing returns, until such a point as they are better off not having pushed past this point.

As you can see in that handy little chart, after a certain point the utopia that we’re all seemingly working towards becomes an unachievable and unprofitable pursuit, and one that we over-invest in when we should be focused on the fundamentals.

For example:

Product / offer catalogue – If you’re a supermarket with 30,000 different SKU lines, you’ve got a pretty good range (although not unlimited) of different offers to put out to customers.  If you’re a company that only sells 1 x SKU to a specific customer cohort, then maybe your offer catalogue is more restrictive and you don’t have the same natural diversity in your range to warrant higher order personalisation.

Your data – The volume / quality of your data may not be sufficient to personalise with sufficient accuracy beyond a certain point.  Potentially from a customer privacy point of view, pushing personalisation past a certain point could create a sense of creepiness, or push sensitive data into more risky workflows and increase security risks.

Production cost / operational overhead of differentiated engagements – Whether it’s dynamic advertising assets, chatbot flows… these things can definitely be automated, but someone still has to care about and maintain each variant in some capacity.  This creates operational complexity in the organisation.  Maybe it’s the organisation of measurement that creates the complexity.  The ‘total cost of ownership’ of each new personalisation feature or experiment needs to be considered holistically though in terms of people/process and tech impacts.

Exposure reach – Interestingly even those figures such as Byron Sharp who have made logical statements around the need for greater consistency in marketing assets to maximise ‘mental availability’ effects, also acknowledge that “Segmentation is fine”. However clearly media exposure and messaging reach plays a role as to whether appropriate frequencies of messaging can be built to create genuine salience / action if scarce moments are overly segmented.

Creative platform / idea – Some ideas are more suited to granular expression, others a better focussed on single-mindedness.  Having a sensitivity to whether personalisation is fighting with this or detracting from it also needs consideration.

Identifying the point at which pursuing deeper personalisation will create a diminishing ROI on these dimensions can require some level of modelling, but it’s important to ensure that any new ideas are validated against these kinds of filtering conditions. 

I should add, that innovation can also change these constraints. Brands like M&M’s, Nike etc. have commercialised the personalisation of their products.  And new technologies continue to change the operational costs of executing scalable personalisation.

The key point is to make choiceful decisions on where you feel your organisation is at right now. 

Do you actually have the right ideas to make this stuff work? Will the personalisation lead to a more creative, more innovative and engaged experience? Or are you setting a false horizon?   

And when we consider personalisation, make sure that you’re confident that you’re going to have something interesting to tell people. There’s no use being able to target all of your customers with a personalised message if the message itself is shit.

As important as the ability to personalise our comms is, we can’t ever forget the importance of having something interesting to say in the first place.

So as we continue along the personalisation journey, let’s just ensure that we think, rather than relying on generic models designed by over-zealous management consultants trying to build burning platforms for us to close a deal to tell us what our end-goal should be.