Licensing Your Data vs Using It Yourself

Once a business realises its archive might be worth something, the question that follows is usually whether it could be sold.

It is the wrong first question, and the answer for most UK SMEs is no. But it is worth understanding why, because the comparison clarifies what your data is actually good for.

This article is part of our guide to the data you already own.

The two propositions

Licensing means selling access to your data, or to insight derived from it, as a product. Someone outside your business pays for what you know.

Internal use means using the same information to improve your own pricing, targeting, delivery or product. Nobody pays you for the data. You simply make more money with it.

They are not exclusive, and they demand very different things of a business.

The honest comparison

	Licensing	Internal use
Who pays	A third party	Nobody, it shows up in margin
Time to first return	12 to 24 months	3 to 6 months
Investment	£50,000+	£10,000 to £30,000
Legal complexity	High	Moderate
Ongoing obligation	Substantial: currency, support, contracts	Low
Who it suits	Businesses with genuinely unique, large, current data	Almost everyone
Typical outcome	A small number succeed well	A reliable margin improvement

Four conditions licensing requires

If you fail any of these, it will not work. Most businesses fail at least two.

The data must be genuinely unique. Not "we have a lot of it". Nobody else can obtain the equivalent. If a buyer can get something adequate elsewhere, more cheaply, they will.

It must be current, and stay current. A dataset is a depreciating asset. Selling access means committing to keep it fresh indefinitely, which is an operational obligation you take on forever. This is the condition that most often kills a promising idea after launch rather than before.

There must be identifiable buyers with budget. Not "this would be interesting to the industry". Named categories of organisation with a reason to pay and a budget line it comes from.

You must be legally free to do it. The hardest one where the data concerns people or clients. Personal data requires a lawful basis for the new purpose, and client confidentiality frequently restricts use more tightly than data protection law does. See can you legally use your own customer data to train AI.

What licensing actually looks like when it works

It is rarely raw data. Almost always it is one of three forms:

A benchmark. Clients compare themselves to an aggregate only you can see. Sold as a subscription or as part of a service. This is the most achievable version for an SME, and it is really a product built on data rather than a data sale.

An index or reference. Regularly published figures the market comes to rely on. High barrier, long build, very durable once established.

Embedded access. Another business builds your data into their product and pays for the feed, which makes you a software supplier with the obligations described in MVP vs full product. Highest revenue potential, and it makes you a supplier with uptime obligations and integration support, which is a different company from the one you currently run.

Why internal use usually wins

Three reasons, and they are not close.

The return is faster and more certain. A pricing analysis delivers something actionable within a quarter. See your pricing history is a dataset, where a £14,000 project produced roughly £71,000 of annual margin.

It requires nothing new from the organisation. No new sales motion, no contracts, no support function, no ongoing currency obligation. You already have customers and you already set prices.

It is a precondition anyway. If your data cannot improve your own decisions, it will not improve a buyer's either. Internal use is the test of whether licensing is even plausible, and it pays for itself while you find out.

The sequence

If licensing genuinely appeals, do it in this order.

Use it internally first and see whether it produces anything. If it does not improve your own decisions, stop.

Turn the insight into a benchmark you give clients, as part of what you already sell. Free, initially. This tests whether anyone values it without requiring anyone to buy anything.

Charge for it once clients ask for it unprompted. That request is the only reliable demand signal.

Only then consider a standalone product or a licensing arrangement.

Businesses that reverse this order build a data product nobody validated, which is one of the more expensive mistakes available.

Key Takeaways

• For most UK SMEs, licensing is the wrong answer and internal use is the right one.
• Licensing requires four conditions: genuinely unique data, an indefinite commitment to keeping it current, identifiable buyers with budget, and legal freedom to do it.
• The currency obligation is what usually kills licensing after launch. A dataset depreciates, and selling access means maintaining it forever.
• Internal use returns faster, costs less, requires nothing new of the organisation, and is a precondition for licensing anyway.
• Sequence it: use internally, give it away to clients, charge when they ask unprompted, then consider a product.

Frequently Asked Questions

Could we sell our customer list?

Almost certainly not lawfully, and you would not want to. It requires a lawful basis and transparency almost no business established at collection, the reputational cost is real, and the money is poor. This is the single most common version of the question and the answer is consistently no.

What about selling anonymised data?

Legally much more workable, and it runs into the commercial conditions instead. Once anonymised, is it still unique and useful enough that someone would pay? Frequently the anonymisation removes exactly what made it valuable.

How much could a benchmark product realistically earn?

For a UK SME, typically tens of thousands rather than hundreds of thousands, and its larger value is usually in winning and retaining the core business rather than in the subscription revenue. Treated as a differentiator it often pays. Treated as a new revenue line it usually disappoints.

Is it worth talking to a data broker?

You can, and set your expectations low. Brokers want scale, currency and clean provenance, and data quality is usually where SME datasets fall down. Most SME datasets fail on at least two, and the conversation is often useful mainly for confirming that internal use is the better route.

---

Wondering whether your data is a product or a margin improvement? Talk to Halo Technology Lab. Our strategy and scoping service will give you a straight answer, and it is usually the less exciting one.