How Bad Is Data Decay, Really? The Math Every Dealer Should See

How Bad Is Data Decay, Really?

A dealer asked me this in a room last month. He had heard the number that a database is 40% wrong after six years and he did not believe it. It sounded too high.

It is not too high. For some fields it is low. Let me walk the math, because once you see it you cannot unsee it, and you will never look at a mail drop the same way again.

The decay rates, field by field

Your database is not a photograph. It is a melting ice cube. Every month, a slice of it stops being true.

Email is the fastest. The most-cited industry benchmark puts decay around 2% a month. North of 20% a year. Past 30% in fast-moving markets. The address that reached your customer in January can be a bounce by December.

Phone numbers churn up to 25% a year. People switch carriers, drop landlines, change numbers.

Households move more than 10% a year. New address, new everything.

Vehicles are the big one. Nearly two-thirds of Americans keep a car five years or less, so past six years roughly 40% of the vehicles in your file are no longer owned by the person your record says owns them. Traded, sold, totaled, gone.

None of these decay in sync. They compound. A record can have the right name, a dead email, an old phone, and a car the customer sold two years ago. Technically it is one record. Functionally it is noise.

What this does to a campaign

Here is where the math gets expensive.

You send 10,000 emails. You feel good. Ten thousand touches. But at a 30%-plus annual decay rate, a big share of that list is already dead on arrival. Realistically you reached maybe 6,000 real inboxes. You paid to send 10,000.

Then the floor stays empty and everyone asks what went wrong with the campaign. Nothing went wrong with the campaign. The campaign was fine. You were marketing to ghosts. Four thousand of them.

You cannot out-create a bad list. The best offer in the world sent to a bounced address is a bounce.

Why a one-time scrub does not fix it

The instinct is to clean the database once and call it done. That is a haircut, not a cure.

Data decays continuously, so it has to be cleaned continuously. A one-time scrub is accurate the day it finishes and starts rotting the next morning at 2.5% a month. Six months later you are most of the way back to where you started.

Clean data is not a project. It is a utility. Like the lights. You do not turn them on once.

The real cost is the decisions

Wasted postage is the small cost. The real cost is that every decision sitting on top of that data is a little bit wrong.

Your equity mining is chasing people who sold the car. Your service reminders are going to dead addresses. Your "loyal customer" segment includes people who moved away and defected two years ago. You are steering with a fogged windshield and wondering why the turns feel late.

Clean, operational-grade data is not a marketing nicety. It is the floor everything else stands on. Get it right and the reports get honest. Get it wrong and every number downstream inherits the rot.

The takeaway

The 40% figure is not scary marketing. It is arithmetic. Email 2.5% a month. Phones 25% a year. Vehicles 40% past six years. It compounds, it never stops, and a one-time clean does not hold.

Stop paying to reach ghosts. Start treating your data like the asset it is, which means cleaning it continuously, on infrastructure you own.

Related on QoreAI

• How Do You Protect Your Data While Using AI?
• What Is a Lakehouse?
• What Is a RAG?
• QoreCloud: your dealership brain