Short answer: original data is worth publishing because it is expensive for a competitor to copy, not because publishing it triggers a signal. Run the test: could a competitor produce this by editing their own page? If yes, it is a statistic, not original data. Most local service businesses already have a good source of it sitting in a spreadsheet, because it is the record of work they have done.
The test for whether data is actually yours
There is a clean test, and it does not require any technical knowledge.
Ask: could a competitor publish this same figure by editing their own page? If your claim is “average boiler replacement cost in Kent is £2,400”, they can copy that number and change the county. If your claim is “we have replaced 340 boilers in Kent since 2019 and 90% are still running after five years”, they cannot — because it is a record of your work, and a number you did not generate is a number you cannot defend.
That distinction is the whole thing. Most published “statistics” on business websites are the first kind, which means they carry no advantage at all beyond the effort of finding them. The second kind is scarce and valuable, and most businesses have it already without having written it down.
Key takeaway: originality is about the source, not the presentation. If you did not measure it, you do not have it.
Where your data already is
Almost nobody needs to run a survey. Look at what a service business accumulates by accident:
- Job records. How many, how long each took, what they cost, what happened afterwards.
- Failure and return rates. What proportion came back within a year, and why. Nobody publishes this.
- The questions you keep being asked. Note them for a month. The frequency ordering is itself the finding.
- What you refuse. How often you decline a job, and on what grounds.
- Timing. When you are actually reachable, how fast you reply, how long a quote takes.
Each of these is a filter on records you already keep. The work is not collecting data, it is noticing what is in it. A practitioner is generally better at this than an analyst, because the interesting observation is usually something that struck you as odd — and running that through someone else’s dataset tends to flatten it. A content audit is a reasonable way to find which of your own records nobody has looked at yet. [4]
How to publish a finding without overclaiming
The difference between a credible finding and an embarrassing one is almost entirely in the framing. Four things make the difference:
- State the scope. How many jobs, over what period, in what area. “From 340 jobs since 2019” is credible; “our customers report 98% satisfaction” is not, because it is unmeasurable and everyone knows it.
- State the limit. If you only track jobs over a year, say so. The caveat is what tells a reader you were careful, and it is the opposite of hedging — it is the evidence of care.
- Avoid percentages where counts are clearer. “9 of our last 40 repairs needed a part replaced within a year” is more informative than “22.5% return rate”, and nobody has to guess your denominator.
- Keep the unflattering findings. A business publishing the job it turned down sounds honest, which is worth more than the omission.
Google’s helpful-content guidance is really the relevant standard here: original, substantial content that helps someone with a real question. [3] A finding you measured and a statistic you found are not equally substantial, and readers can tell.
What happens next, and what does not
There is no originality signal. Nobody’s system rewards a page for containing data that appears nowhere else, and the specific claim that publishing a statistic forces an AI to attribute it to you does not hold up — an assistant summarising several sources will simply state the figure without naming anyone. The research on making documents more quotable is about clarity and evidence attached to claims, not about uniqueness. [1]
What does happen is more ordinary and more reliable:
- People link to it. Not because of a signal, but because a specific number is easier to write about than a vague one. This is how most original research actually earns its return, and it takes months rather than weeks.
- It gets cited by people, which is worth more than a link and is described in brand reputation.
- It is defensible. When a competitor publishes a restated version, you can point at the original — and there is evidence that this kind of sourcing tends to get linked more than the restatement. [2]
Key takeaway: the citation follows the data. Publish something competitors cannot duplicate and the rest is ordinary compounding.
| Claim | Could a competitor publish it? | Value |
|---|---|---|
| ”Industry average job takes 2 weeks” | Yes, by editing their page | Near zero |
| ”We have done 340 jobs since 2019” | No | Real, and it compounds |
| ”98% of our customers are happy” | Yes, and nobody believes it | Negative |
| ”9 of our last 40 repairs needed a part within a year” | No | High, and it reads as honest |
| ”We turn down about one job in five on price” | No | High, and it builds trust |
What to do first
- Open your job records and write down five numbers you could state with a scope and a date. That is your whole dataset.
- Pick the one that contradicts the usual industry claim.
- Write it up with the scope, the limit, and one unfladding detail.
- Publish it, then link it from the service page it describes, and write it the way you would explain it to a customer. [5]
- Do nothing clever. Update it next year with a new figure.
How this connects
- How-to guides covers the other uncopyable asset: your method.
- A justification web covers attaching evidence to every claim, including these.
- Building trust online covers what happens when people start citing you.
- Freshness covers updating the figure when the next year lands.
