Hot Glass Looks Exactly Like Cold Glass

Glassblowers learn that the hard way. Business data is the same. A list of good contacts and a list of bad ones look identical on screen: the same names, the same job titles, the same tidy phone numbers. We've said it for almost 35 years. Good data is indistinguishable from bad data until you use it.

So buyers judge on the one thing they can see: the price.

Economists have a name for what that does to a market. In 1970 George Akerlof showed it with used cars. If buyers can't tell a good car from a bad one, "good cars and bad cars must still sell at the same price". Owners of good cars can't get what they're worth, so they stop selling. In his words, "the 'bad' cars tend to drive out the good". He called the bad cars lemons, and the idea won him a Nobel Prize.

Business data went the same way. When price is all a buyer can see, the suppliers who pay to keep data right can't be paid for it, and most of them have gone. What's left is a large, cheap market in data that looks fine on screen, and a few suppliers still doing the work. Corpdata is one of them.

Some bad data gives itself away. Emails bounce, phones ring out, post comes back. It's annoying, but at least you can see it.

The expensive faults stay hidden. The contact who left last year, whose emails now go to a colleague who has never heard of you and reports you as spam. The finance director who now runs facilities. The business that looks right on paper but will never buy what you sell. Nothing bounces. The campaign just goes quiet.

And when a campaign goes quiet, nobody blames the list. They blame the channel: "We tried direct marketing. It doesn't work for us." That verdict can last for years, long after anyone remembers where the data came from. Akerlof saw this too. The real cost of a bad product, he wrote, includes "the loss incurred from driving legitimate business out of existence".

Then there's the risk you can't hand back. Under UK law, the business that uses the data is responsible for how it's used, whatever the seller promised. The regulator is blunt: "Simply accepting a data broker's assurances that the data they are supplying is compliant is not enough."

"Verified" sounds like one thing. It's at least four:

1. The contact works. The email doesn't bounce; the number rings.
2. It belongs to the person named.
3. They still do that job, at that business.
4. You're allowed to use it for what you plan to do.

Proving one doesn't prove the others. A check that a mobile is switched on tells you about the number, not the person holding it. An inbox that accepts email can belong to someone who left last spring. A lot of "verified" data has only passed the first test.

Mobile numbers show why this matters in the UK. The rules for marketing to a number depend on who pays the bill. If a company pays, it's a business line. If the person pays, the law treats it as personal, even if they only use it for work. Nothing in the number tells you which, and no network check can. When you can't tell, the regulator says to treat it as personal. Mobiles can be registered with the Telephone Preference Service too.

So what happens if you use it? If the number is personal, you needed the person's consent before you texted it. If it's on the Telephone Preference Service, you shouldn't have called it. Complaints are how the regulator finds you, and fines for unlawful marketing calls and texts can now reach £17.5 million or 4% of turnover. As card 2 showed, it's the business that used the data that pays, not the one that sold it. And even if nobody complains, you've just rung a decision-maker on their own phone, uninvited. Not the best first impression.

That's why the contact detail that looks most valuable - a decision-maker's mobile - can be the riskiest one you can buy in the UK. A lot of data built for markets with different rules leads with exactly that: a mobile number, "verified" because it rings.

If you can't see quality, how do you buy it? Akerlof's answer was to look at who carries the risk: guarantees, a name with a record to protect, a buyer who is free to walk away.

Economists have since tested it. In a 2011 experiment with 936 people, markets like this worked when the seller was on the hook for the result. Letting sellers build a reputation "has little influence". More competition "drives down prices", but it didn't make the outcomes any better.

A European data regulator said much the same this year. Fining two businesses that had bought contact data from a global platform, it criticised buyers who rely on "merely formal guarantees of compliance", trusting a platform's reputation or its "appearance of professionalism".

So ask any supplier four questions. Here are our answers.

• How fresh is it? On average, every record was last checked 94 days ago. We keep checking, and we start with the records most likely to have changed.
• What happens if it's wrong? Our written 2-for-1 Goneaway Guarantee. The checking is the point; the guarantee is the proof we mean it.
• How long have you been doing this? Almost 35 years. We started in February 1992.
• Can you show me the compliance answers? Yes. Our answers to the regulator's due-diligence questions are published for anyone to read.

Don't take our word for it. Ask us how it's measured.

Then ask everyone else the same questions, and see how many UK suppliers still put their own money behind their data.

Tell us who you need to reach and what a good result looks like. We'll find the data most likely to deliver it, and we'll tell you plainly if a target is hard to reach. Before you buy, we'll show you how fresh it is.