Your Site Says 100%. Can You Show the Number?

Somewhere on your website there is a sentence you did not write carefully. It says you are the best, or the leading provider, or that you have a 100% success rate. Nobody has ever challenged it, so it has sat there through two redesigns. Then a buyer compares you against three rivals who say the same thing, and the sentence does no work at all.

A claim with its workings shown beats a bigger claim with nothing behind it. This is a decision, not a rule: for each claim on your site you can keep it and publish the basis, soften it to something you can stand behind, or delete it. We would take the first option where the evidence exists and the third where it does not.

The claim nobody asked you to prove

Here is what we measured this week. On 27 October 2026 we read 30 pages across 12 Singapore music-teaching businesses for a separate study of how those sites publish fees and policies. While we were in the text, we counted results claims.

Four of the 12 made a numerical claim about their own results. Two said they had a 100% exam pass rate. Two gave a count of students taught, one in the hundreds and one above three hundred.

None of the four said how many students the number covers, or over what period, or which exam board. One came closest, with the words "to date" before the pass-rate claim, which tells a reader the period is open-ended and nothing else.

We are not saying any of those four businesses did anything wrong, and we name none of them. A 100% pass rate may well be exactly true. The point is narrower: as published, the claim cannot be checked by the person deciding whether to spend a few hundred dollars a month, and an unverifiable number has to compete against every rival's unverifiable number.

That is the whole problem. A claim everybody can make is not a claim.

Why the sentence is there in the first place

Because buyers search for it. In the same run we pulled 79 Google Autocomplete suggestions for three Singapore music-lesson seeds. Thirteen of the 79 used best, good, review, Reddit or "worth it". Google's own documentation says that "autocomplete predictions reflect real searches that have been done on Google", and we read that page on 27 October 2026.

So the instinct is sound. People genuinely search for the best. The mistake is answering a comparison question by asserting the conclusion. A buyer searching "best piano teacher singapore" is not looking for a business that says it is the best; they are looking for something that lets them decide. A dozen pages all asserting that they are the best give them nothing to decide with, so they go and read somebody else's round-up of your trade instead, in a search you ought to own.

What the rules actually say in Singapore

Two documents matter, and neither is as dramatic as owners fear.

The advertising side is self-regulatory. The Advertising Standards Authority of Singapore publishes the Singapore Code of Advertising Practice, and its page on the Code states that "the basic premise of the SCAP is that all advertisements should be" legal, decent, honest and truthful. The download on that page is labelled SCAP 2008, with an updated chapter introduced in January 2019. We read the ASAS page on the Code on 27 October 2026.

The consumer-protection side is statutory. The competition and consumer commission's page on the Consumer Protection (Fair Trading) Act says the Act was "enacted to protect consumers against unfair practices", that section 4 "defines what is an 'unfair practice'", and that its Second Schedule contains "a non-exhaustive list of unfair practices". That page records a last update of 20 July 2026 and we read it on 27 October 2026.

What neither page gives you is a word list. There is no register of banned adjectives for ordinary businesses. Some trades do have one: a Singapore clinic works under advertising regulations with specific restrictions on laudatory language, which we covered in the clinic advertising rules and the words to avoid. If you are in healthcare, legal or financial services, your regulator outranks everything in this article.

For everyone else the practical test is the one in the SCAP premise: is the claim honest and truthful, and could you show someone why. This is marketing guidance, not legal advice, and we are not your lawyers.

The three choices

Take one claim at a time and pick one of three.

Keep it and publish the basis. Swap "a 100% pass rate" for the sentence your own records support. If the register shows every student you entered for practical exams over the last three years passed, say how many that was and name the years. The claim gets narrower and stronger at the same time, because now it can be checked. Dates and denominators are what turn a boast into evidence.

Soften it to something you can stand behind. "The best piano school in Singapore" becomes "one-to-one lessons with diploma-qualified teachers, at a stated price for a stated lesson length, five minutes from the MRT". Less flattering, far more useful, and it answers the question the buyer actually had.

Delete it. If you cannot say where the number came from and you are not willing to count, the sentence is doing nothing except making your page look like the other eleven. Cutting it costs you nothing you can measure.

In our view most owners should run the second option across the whole site and the first option on exactly one claim, the one they can prove best. One well-sourced number does more than six adjectives.

What each choice costs

Nothing here is free, and anyone telling you otherwise is selling.

Publishing the basis caps the claim. A sentence naming how many students you entered invites the obvious follow-up about how many you did not enter, and you need an answer. It also dates your page, so it needs reviewing each year, and a stale proof line is worse than none.

Softening loses the emotional punch, and if your competitors keep shouting you will feel quieter for a while. In a shortlist where the buyer compares pages side by side, specific usually wins; in a quick scan it sometimes does not.

Deleting is cheapest, but it leaves an absence. If the claim was carrying real weight on your homepage, replace it with a fact rather than white space.

And counting takes work you may not want to do. Pulling three years of exam entries out of a spreadsheet is a Saturday, and some owners will find the real figure is not 100% after all. In our view the real figure, with the method shown beside it, does more for a buyer than 100% with nothing behind it. You have to be willing to look.

A ten-minute audit of your own claims

1. Search your site for the words best, leading, top, trusted, premier, award-winning, number one and guaranteed. Every hit is a claim.
2. Search for the figures: a per cent sign, "pass rate", "+ customers", "+ students", "years of experience".
3. Beside each one, write the source in a single line: the count, the period, and where the record lives. If you cannot write that line in a minute, the claim fails.
4. Rewrite or cut everything that failed, in one sitting.
5. Pick the single claim with the best record behind it and put the basis on the page, with a date.

Then diarise it for twelve months. A dated claim that has gone stale is the version a careful buyer notices.

Where we failed this test ourselves

Our own trade is the worst offender, and we include ourselves in that. When we audited 60 Singapore agency pages selling AI-search work in October, not one showed dated proof of a result. Agencies sell measurement and publish adjectives.

We also keep a ledger of our own experiments rather than only the wins, and when we ran the numbers, 0 of 34 changes proved statistically significant. That is an uncomfortable sentence to leave on a sales website. It is also the only honest version, and it is the standard we are asking you to hold your own page to.

So here is the concession, plainly: if the only thing wrong with your website is three unprovable adjectives, you do not need an agency. You need an hour and a willingness to delete your own writing.

One next step

If you would rather have someone else read your claims the way a sceptical buyer does, send us your homepage and your main service page. We will list every claim we find, say which ones we could not verify from your own site, and tell you what we would publish instead. That starts as a free check through our contact page.

If you want the work done with the rest of your search marketing, our SEO, AEO and GEO programme is S$750/month with 20 articles a month. We will not promise you rankings or AI citations in return. AI platforms choose their own answers.

Claim counts on this page describe 12 Singapore music-teaching websites read on 27 October 2026. No business is named and nothing here is a finding about any company's compliance.

FAQ

Is it against the rules to say my business is the best in Singapore?

There is no general list of banned words for ordinary businesses. The ASAS page on the Singapore Code of Advertising Practice says advertisements should be legal, decent, honest and truthful, and the consumer commission's page on the Consumer Protection (Fair Trading) Act points to a non-exhaustive list of unfair practices. Regulated trades such as healthcare have stricter rules. This is not legal advice.

What makes a performance claim credible on a website?

A denominator and a date. A sentence naming how many students you entered, over which years, and how many passed can be checked; "100% pass rate" cannot. In the 12 Singapore sites we read on 27 October 2026, four published a results number and none stated the cohort size or the period it covered.

Should I delete a claim I cannot prove?

Usually yes, unless you are willing to count. A claim every competitor also makes adds nothing to a shortlist, and the space is better spent on a specific fact: your price, your lesson length, your qualification or your location. Deleting costs you nothing measurable.

Do testimonials count as proof of a result?

They are evidence of one customer's experience, not of a rate. Two of the sites we read showed parents describing exam distinctions, which is useful social proof but not a pass rate. Keep testimonials, and keep them separate from any number you publish about results.

How often should a proof claim be rechecked?

Once a year, with the date on the page. A dated claim shows a careful buyer that someone is maintaining the site; an undated claim from four years ago raises the question of what has changed since. Put a reminder in the calendar when you publish it.