Valuing data in M&A

How should you think about the value of the data when you buy a company?

Before we dive into content this week, I wanted to say a quick hello to the 20 new subscribers since the last instalment. It’s a pleasure to have you join us and I hope you will find something valuable here for you this week and every week. Also a periodic reminder to exisiting long time subscribers to share the newsletter with others you know who might enjoy it. Many thank yous. OK, on with the show.

If you hadn’t previously been aware how valuable (certain types of) data can be, 2024 was probably the year when that fact jumped up and smacked you in the face.

But data is not something you’d want to blindly value by the terabyte. At scale it is expensive to store and secure. And when someone steals it, that can be enormously impactful both to your company and to your customers.

How then do you think about the value of data when you are buying or investing in a company? This is more art than science and there is a whole lot of ‘its depends’ but below are some dimensions to consider.

Coherence

I’m a long time user of the Google suite of lifestyle products - mail, calendar, docs, maps - and I find it about equal parts helpful and creepy that my calendar appointment locations show up in Maps and the idiosyncratic spellings of names of friends and family follow me around the various places I write prose. Google isn’t unique in this, Microsoft and Apple would I’m sure do the same if I lived more substantially in their semi-walled gardens. But this level of coherence is actually really rare when you measure along the ‘number of companies who can do it’ axis rather than along the ‘footprint of those companies on our collective web based experience’ axis.

Basically, these ‘coherent data’ companies have figured out how to have one version of Kendra and to link together all the actions that Kendra takes in various products across their suite. The detail of how isn’t important for this discussion (as many variants as you have architects 😎) but it is remarkably hard to do and takes both investment and persistence.

Used well, data coherence is the bedrock of a standout customer experience.

Unfortunately our collective view of how feasible it is to be a coherent data company and offer this great customer experience uplift, is distorted by the enormous footprint of the few companies who have invested heavily to do it well, because their business model and their scale both demand it and enable it. Many non technical executives and board members are inadvertently wrong footed by their personal consumer apps experience into thinking that this level of coherence is the norm and hence easy/cheap to achieve. Don’t fall prey to that understandable but incorrect assumption.

Check to see whether the company you are considering has the pervasive customer knowledge that comes from coherence in the user data. The best way to do this is to use the product / product set pretty extensively. No one is going to tell you that they have poor data coherence after all. Unless the company is very small or very ‘data forward’ it probably doesn’t. Are you unconsciously valuing in a trait/capability that doesn’t actually exist?

Additionally, if you are considering an acquisition, don’t overlook the integration challenge of maintaining and extending any existing coherence of experience into your own product set. It’s fine to make the decision not to integrate at a data level (the smart move in many cases as the work is really tough, expensive and largely invisible to buyers as opposed to end users) but be eyes open about the impact of that decision on the end user experience and make sure you’re not buying one reality while imagining, and hence pricing for, a different one.

Uniqueness

Having access to data that no one else does has been and continues to be an advantage worth pursuing. That data might be tangible e.g. the purchasing habits of a particular segment of the consumer or business market or it might be a byproduct of your product usage e.g. granular, individual watching/listening data for Netflix/Spotify.

Being alive to this dimension can help you price competitively in markets where the value of unique data is understood and, perhaps, to identify a bargain in areas where there has traditionally been less focus on the ‘digital breadcrumbs’ that result from user actions inside software products.

But be careful to consider this Uniqueness dimension is tandem with the Social License dimension discussed below.

Freshness / Relevance

One nuance I’ve seen go astray in valuations is failing to distinguish between data that is gathered in ‘one time capture’ and data that has a built in ‘update mechanism’ in the course of product usage.

For instance consider capturing either household composition or total income in the initial application for a credit product. That is one time capture and the data will have limited usefulness (even assuming you have the social license to use it for purposes other than credit decisioning).

I see teams make this mistake a lot. Always consider the natural update cycle and compare and contrast the expected freshness of a data point of interest relative to the ‘decay curve’ of its relevance in any intended use cases.

Social license

Let’s be clear, this has always been an important dimension. You shouldn’t do things with people’s data that you wouldn’t happily tell them about. That’s creepy and it’s a crappy thing to do.

However there has been a big rise in the prominence of this dimension recently because of an increased awareness across society of the value of personal data, the stratospheric valuations of generative AI based companies, and the swirling, politically charged negotiations about AI regulation.

Note that I deliberately talk about social license rather than regulation because while meeting relevant regulations is important, I also believe there can be a lot of unintended consequences from a one eyed focus on regulation. Rather, step back and consider what usage you would feel comfortable being associated with, as a human and as a brand.

When you are thinking about acquisition, it’s important to consider whether the existing social license for data use that your target has will realistically transfer to you and your intended use. Was this an obvious use that a reasonable person would have expected when first signing up to use the original product? If you need to update your terms of service for the broader usage you are now considering, will a substantial number of users happily agree or will you uncover a whole lot of sleeping dogs who, given the wake up nudge, will withdraw their consent and usage?

Essentially my advice is here is don’t overpay for a data asset that will evaporate on closer examination. Again, a mistake I’ve seen too many teams make.

To recap

The quantitative, data loving scientist within me is always delighted to see data as a significant consideration in a valuation conversation. But do yourself a favour and do it mindfully.

• Coherence

• Uniqueness

• Freshness / relevance

• Social license

I hope you find this helpful. If you’re deep in the data valuation game, I loved to hear what additional dimensions you use to help you reach a number you feel good about. Please add a comment or ping me back.

And if you’re a founder or building a young company, consider how you might add value to your company in the eyes of investors and acquirers by paying attention to and having a clear story about your use of and build for data.

It’s super chilly here in Melbourne now and we’re scheduled, finally, for a decent amount of rain. So likely an inside weekend for me. Wherever your weekend finds you, I hope you can find time for a break in close proximity to nature and an unwind. Maybe over a great cup of tea. Enjoy!

Photo credit to Aleksandr Popov on Unsplash