Every list contains errors. Vendors imply otherwise, investors assume otherwise, and the gap between the two costs real money in ways that are difficult to see.
The problem is not that data is imperfect. It is that an unmeasured error rate makes every other number in your marketing unreliable.
What Actually Goes Wrong in a List
Accuracy is the assumption sitting underneath every other decision in lists and data for real estate investors.
Stale ownership. The property sold and the record has not caught up. Assessor data updates on its own schedule, sometimes annually.
Wrong mailing address. The owner moved and never updated the tax record, so your mail reaches a previous address.
Deceased owners. More common than investors expect, and it produces mail addressed to someone who has died arriving at a grieving household.
Entity confusion. The record shows an LLC and the person who would act on your letter never sees it.
Wrong phone numbers. The largest category, because skip-traced contact data is probabilistic. A share of returned numbers belong to someone else entirely.
Duplicates. The same owner appearing several times under name variations, so they receive four copies of your letter and conclude you are careless.
Wrong property characteristics. Square footage, beds and condition indicators that do not match reality, which affects your filtering and your initial valuation.
The Direct Costs
Easy to calculate and rarely calculated.
Mail to a wrong address is the postage plus the print, spent on nothing. At any volume that is a meaningful line.
Calls to wrong numbers are the caller's time plus, more seriously, contacts with people who did not want them and who may report them, which carries consequences beyond wasted minutes, per calling and texting rules.
Skip tracing spend on records that were never going to be workable.
And staff time working records that could not have produced anything, which at any scale exceeds the data cost itself.
The Indirect Cost That Matters More
Errors do not only waste spend. They corrupt your measurement.
If a third of a list is wrong, your response rate is calculated against a denominator that includes records nobody could have responded from. Your cost per lead is inflated by spend on nothing. Your conclusion that a niche does not work may be a conclusion about a data source.
That is the expensive version, because it produces a strategic decision based on an artifact. An investor who abandons a genuinely good niche because the vendor's data for it was poor has lost considerably more than the mail cost.
Which is why the error rate belongs in the analysis rather than in the background, worked through in why small sample marketing numbers mislead.
How to Actually Measure It
An afternoon, and almost nobody does it.
Pull thirty records at random from the list. Not the first thirty, which are frequently sorted in a way that biases the sample.
For each, verify the owner against the county record directly, check the mailing address, and confirm the property still appears to be owned by that party.
For a sample of the skip-traced numbers, call them and note whether the person answering is who the record says.
Then calculate. Thirty records gives you a rough rate, which is enough to decide whether to spend against the list and enough to compare two providers meaningfully.
Do this before every significant campaign with a new source, and periodically with an existing one, since accuracy varies by county and drifts over time.
What an Acceptable Rate Looks Like
There is no universal number and there are useful reference points.
Ownership and address data from county sources is generally the most reliable, since it is the source rather than a copy of one.
Vendor property data is usually good and varies by county, because vendors aggregate from the same county sources with different refresh schedules.
Skip-traced phone data is materially less reliable than any of the above, and that is inherent to what it is rather than a failing of a particular provider, detailed in why your hit rate is low.
The practical standard: know your rate rather than benchmark it. A list you know is eighty percent accurate can be worked profitably. A list you assume is perfect and is not will produce conclusions you cannot trust.
Reducing the Error Rate
Several steps, in order of return.
Deduplicate before anything else. The cheapest fix and it removes both waste and the impression of carelessness.
Verify against county records for high-value records. Not for ten thousand, for the two hundred you are about to work intensively.
Use the most recent source available. A vendor refreshing monthly beats one refreshing annually, and the county beats both.
Cross-reference two sources where the stakes justify it. Agreement between independent sources is a reasonable proxy for accuracy.
Remove records that prove wrong. Feed the result of every wrong number and returned letter back into the file, which most investors never do, so the same errors are worked repeatedly.
Verifying Without Doing It by Hand
Some of the checking can be mechanical, which makes it more likely to happen.
Address validation against postal data catches malformed and undeliverable addresses before you print anything, and it is inexpensive.
Deduplication on a normalized address or parcel identifier catches the same owner appearing several times.
Cross-referencing two sources on ownership flags records where they disagree, which is a good proxy for the ones worth checking manually.
Deceased-indicator screening is offered by several data providers and is worth applying to any list where the niche makes it likely.
None of that replaces the thirty-record manual sample, because automated checks confirm internal consistency rather than truth. They reduce how many records need the manual look, set out in AI for data enrichment.
The Feedback Loop Nobody Builds
The single highest-return habit in this area.
Every returned letter, every wrong number, every conversation revealing the property sold two years ago is information about your list. Recorded, it improves the file. Discarded, the same errors are paid for again next quarter.
The mechanism is a status field on every record and a rule that whoever encounters the error updates it in the moment rather than mentioning it.
Over a year that produces a house list materially cleaner than anything you could buy, which is one of the reasons a worked list becomes more valuable than a fresh one, worked through in managing lists as you scale.
When the Data Is Not the Problem
Blaming the data is convenient, so this deserves a caution.
A poor response rate is more often a message problem, a targeting problem or a follow-up problem than a data problem. Investors switch providers hoping for a different outcome and get the same result at a different price, per why your hit rate is low.
The way to tell is the sample. If thirty verified records show the data is broadly right, the problem is downstream and switching vendors will not touch it.
That is the actual value of measuring accuracy: not only to fix the data, but to know when the data is fine so you stop looking there.
The Errors With Consequences Beyond Money
Some data problems cost more than the postage.
Mail to a deceased owner. Arriving at a household in the weeks after a death, addressed to the person who died. Common with probate-adjacent lists and genuinely upsetting to receive, as in talking to sellers in difficult circumstances.
Repeated contact with someone who asked you to stop. Usually a suppression failure rather than a data failure, and it looks identical from the recipient's side, explored in suppression lists.
Calling a number that belongs to someone unrelated. A wrong-number contact is a call to a person who never had any connection to your business, which carries its own exposure.
Contacting a property owner who sold years ago. Harmless and it signals that you are working from data nobody checked, which reaches your reputation in a small market.
These are the reasons accuracy is a conduct question rather than only an efficiency one, and they argue for verification on any list touching sensitive circumstances.
The Habit Worth Adopting
Thirty records, verified by hand, before spending against any new list.
It costs an afternoon and it tells you whether the campaign you are about to fund is being measured against a real denominator. Every investor who has done it once for a list they assumed was clean has found something, and the ones who have never done it are making decisions on numbers they cannot see the shape of.