The Data Decay Problem: Why Your Best Prospects Disappear From Your Database

Decay is not uniform. B2B contact data rots at around 2.1% a month [1], but it hits your highest-value contacts hardest — because the senior, in-demand people worth targeting most are the ones most likely to move. Your database loses its best records first.

Data decay is usually discussed as a flat percentage: some share of records go bad each year, you top up the losses, life goes on. That framing is comforting and wrong in a way that costs real money. Decay is not uniform across a database. It is concentrated — and it concentrates on precisely the records you can least afford to lose.

The uncomfortable insight is that the people worth targeting most are the people most likely to move. Senior operators get poached. High performers get promoted into new titles. Ambitious directors leave to found something. Movement correlates with value, which means the records a rep is most excited to work are disproportionately the ones pointing at a role the person no longer holds. This article is about why that happens, how to find the damage, and why it forces a specific kind of refresh strategy — one weighted by value, not spread evenly.

1. What data decay actually is

Decay is the gap between what your database says and what is currently true, widening over time without anyone touching the file. A record is accurate the day you capture it. Then reality moves and your copy does not. The record does not announce that it has gone stale; it keeps presenting the same title, company, and location it had at capture, looking exactly as trustworthy as a record that happens to still be correct.

At the aggregate level the rate is well benchmarked: B2B contact data decays roughly 2.1% per month, compounding to about 22-30% a year at baseline and higher — toward 70% a year — when you track several volatile fields at once [1]. That is the headline number most teams plan around. The problem is that planning around the average assumes the losses are spread evenly, and they are not.

2. Why records go bad: the decay chain

The mechanism is a simple chain that runs on repeat. A record is created and is accurate. The person then changes jobs — and across the US workforce, median tenure is now about 3.9 years, the shortest since 2002, with roughly a quarter of workers changing jobs in a given year [2]. The company they were at gets acquired or rebranded. Your database, meanwhile, stays frozen at capture. A campaign goes out against the old record, and only then — at the most expensive possible moment — does a bounce or a confused reply reveal the record was wrong.

The cost is not in the decay itself; it is in discovering it downstream, after you have already spent research time, QA time, and sender reputation acting on it. The chain is cheap to prevent at the collection layer and expensive to discover at the campaign layer, and most databases are structured so that discovery only ever happens at the campaign layer.

3. Which fields decay, and how fast

Not all fields rot at the same speed, and the ranking matters because the fastest-decaying field is the one most agencies target on. Job title changes most often — some benchmarks report a majority of titles changing within a single year [1][3]. Company tracks the job-change rate close behind. Work email decays around 3.6% per month because it usually dies with the job move [1]. Location changes slowly. And ICP fit is derivative: it moves whenever title or company moves, which means your segmentation silently degrades every time a contact gets promoted.

Field	Primary cause of change	Relative decay speed
Job title	Promotion, lateral move, reorg	Fastest — majority can change within a year [1][3]
Company	Job move, acquisition, rebrand	Fast — tracks job-change rate [2]
Work email	Job move, domain change	~3.6% per month [1]
ICP fit	Any title or company change	Moves with title/company
Location	Relocation, office move	Slow

4. Why your highest-value records go first

Here is where the uniform-decay assumption breaks. The contacts you most want to reach are usually mid-career high performers — the VP who gets recruited away, the director who gets promoted into a new title, the operator who leaves to start something. Those are the same traits that predict movement. The in-demand and the ambitious change seats more often than the average employee, so the records at the top of your value ranking decay faster than the database-wide average.

Flip it around and the effect is just as sharp: the records that stay accurate longest tend to be the lower-movement, lower-value ones — people in stable roles nobody is trying to poach. So an un-refreshed database does not degrade evenly toward mediocrity. It actively loses its best contacts first and retains its least valuable ones longest. Over a year, the composition of your usable list shifts away from the accounts you actually want.

5. The compounding cost: why value-weighted decay is worse than it looks

This makes decay more expensive than the headline percentage implies. Suppose 20% of a database is stale. If that 20% were random, you would have lost 20% of the database's value. But if the stale 20% is skewed toward your highest-value contacts, you have lost far more than 20% of the value — because the dead records are the ones that would have converted best. The percentage understates the damage whenever decay is value-weighted, and decay is almost always value-weighted.

There is a second-order cost too. A rep burning outreach on decayed high-value records is spending your most limited resources — attention and sender reputation — on your worst odds. Those stale sends drive bounces, and practitioners treat bounce rates under 2% as safe, 2-4% as a warning band, and above 4% as the point where sender reputation degrades faster than warmup can repair it [4]. So the records most likely to be wrong are also the ones a rep is most motivated to hit hard, which is exactly the combination that damages deliverability for every future campaign.

6. How to identify stale records before they cost you

You cannot refresh by value if you cannot see staleness, so the first move is to make both measurable. Store a "last verified against source" date on every record — distinct from the creation date, which tells you nothing about current accuracy. Then layer a value signal on top: segment membership, account tier, or active-campaign status. The two together let you rank records on a simple grid of important-versus-old.

A lightweight way to confirm the value-weighted pattern in your own data: pull a random sample from your highest-value segment and a random sample from a low-value segment, check each record against the live profile, and compare the freshness rates. In most databases the high-value sample is noticeably staler — which is the whole thesis of this piece, demonstrated on your own file rather than taken on faith.

7. Refreshing by value, not uniformly

The strategic implication is direct: do not refresh uniformly, refresh by value. Re-collect current data for your highest-value and highest-volatility segments more frequently, because that is where decay does the most damage and where fresh data pays back fastest. Lower-value, lower-movement segments earn a longer interval. The goal is not to re-verify everything at the same rate — that just spreads scarce refresh capacity evenly over records of wildly different worth — it is to concentrate refresh where it protects the most value.

Mechanically, re-collection should key to the stable LinkedIn profile URL so the current title, company, and location overwrite the existing record in place rather than spawning a duplicate. That preserves your CRM wiring and history while bringing the content back in sync with what the source currently shows. The limit is honest: re-collection keeps you aligned with what a profile currently displays; it cannot update a detail the person never changed, and it does not generate personal emails or phone numbers. What it does is stop your best records from quietly rotting at the top of the list.

8. From periodic cleanup to continuous, prioritized refresh

Done once a quarter, value-weighted refresh is already better than a blind full-database cleanup. Done continuously, it becomes a background property of the database: a rolling process that always re-collects the records most due — the oldest within your highest-value segments — and keeps the composition of your usable list aligned with the accounts you actually sell to. Decay never stops running, so the only durable answer is a refresh process that never stops either, pointed first at the records worth the most.

The reframe is the point. Treated as a flat percentage, data decay looks like a tolerable cost of doing business. Seen as a value-weighted process that strips out your best contacts first, it looks like what it is — a slow erosion of exactly the asset your agency sells — and it demands a refresh strategy built around value rather than volume.