How to test a higher rate on new clients first
You do not learn your price from advice, from a calculator, or from watching what competitors charge. You learn it from quotes - from the rate real clients accept or refuse when real work is in front of them. So run the learning as an experiment: pick a higher rate, quote it to new enquiries only, for a fixed window, while your existing clients stay exactly where they are. No announcement, no risk to the relationships that feed you - and at the end of the window you own something no amount of courage can substitute for: your own results. This guide covers the design, the honesty rules that keep the test clean, how to run it, and how to read what it tells you when it ends.
1. Design the test like an experiment
A useful test changes one thing. Here the variable is a single number: the rate new enquiries hear. Everything else - the work you offer, the proposal, the terms - stays exactly as it was, so the answers you collect are answers about price, not about a new pitch.
Size the step so it matters. As an illustration, worked on the rate this site uses in its examples: a freelancer charging $95 an hour tests $110 - a step of $15 on $95, which is +15.8%. A step too small to feel proves nothing; a step you cannot say out loud without flinching tests your nerve instead of your market.
Give the test a fixed window, and close it by the rule you set at the start: the next ten quotes, or six weeks, whichever ends second. Both halves of that rule earn their place. Ten quotes is a run, not an anecdote - and the six-week floor stops one busy fortnight of enquiries from ending the test before the pattern has had time to show itself. Small samples lie fast, and the first two answers, in either direction, are the noisiest data you will collect.
Every experiment needs its control, and yours is the recent past: your win pattern at the old rate - how many of the last ten or twenty comparable enquiries said yes, and what the work was. You probably already hold most of it in the sheet where you track leads in a spreadsheet: the quotes, their outcomes, their reasons. Read it before the window opens and write the pattern down. Without the before picture, the after picture is just a mood.
The test works the same way if you quote fixed fees rather than hours. The variable is still one number - the rate inside the fee. Build each test quote from $110 instead of $95, with the same hours estimate you would honestly have used before, so the fee moves only because the rate moved. Change the hours at the same time and you have changed two variables, and the answers stop being about price.
Finally, decide in advance what counts as a result. A yes at the new rate is one, obviously. A no is only half a result: it becomes data when it carries its reason - price, timing, fit, silence - noted the way a project retrospective records lessons while they are still fresh. "Lost ten quotes" teaches nothing. "Lost four: two on timing, one on fit, one who named the price" is a finding you can act on.
2. The honesty rules
The test rate is a real price, quoted in earnest. If a client says yes at $110, then $110 is your price for that job - quoted, agreed and invoiced at $110, with no wobble afterwards. The experiment works precisely because every quote in the window is genuine: you are not simulating a market, you are standing in one.
Charging a new client more than an existing one is ordinary commerce, not a trick. Every shop you buy from has re-priced since your first purchase, and nobody calls the corner cafe dishonest for it. Prices move over time; what honesty requires is narrower - that the price quoted is the price charged, to that client, for that job, at that moment.
Two moves do cross the line, and both are worth naming. The first is quoting a price you never intend to charge: an inflated figure invented so a "discount" can land back at your real rate. That is theatre, and clients who spot it leave. The second is lying about the reason. There is no need to claim "my rates have gone up for everyone" when they have not. If anyone asks, the clean line is also the true one: "My current rate is $110."
And one boundary is absolute: existing clients are not the experiment. They never meet the test rate mid-relationship, and their own rise, when it comes, is a separate and announced decision - notice, a date, one short email - handled the way how to raise your rates sets out, or, where the boundary is a renewal, the way the renewal renegotiation guide handles it. The test exists to protect those relationships while you learn, not to smuggle a rise past them.
3. Running it
Put the test rate where new enquiries will meet it: on your rate card, updated once, in the same sitting as the window opens. One card, one current number - the old rate now lives only inside existing agreements, where it belongs. A card still showing $95 while you quote $110 is not a safety net; it is the old price quoting itself on your behalf.
The proposal carries the number the same way it carried the old one: stated plainly, in its place, without apology or decoration - the proposal template needs no new section for a new figure. Do not flag the rate as new, experimental or temporary. A client cannot fairly evaluate a price you have already apologised for, and "testing" language invites exactly the negotiation the experiment was designed to measure past.
Track every quote in the window in the same lead sheet: the date, the work, the rate quoted, the outcome, and the reason where there is one. The sheet is the lab notebook. Ten rows is the whole experiment, and keeping it costs a minute a quote.
Some test clients will negotiate; that is not the experiment failing, it is the experiment working - negotiation is one of the answers a market gives. Hold the number the way you would hold any honest price: trade scope, timing or terms if the budget is fixed, but do not quietly return to $95 for one likeable client and count the quote as a test result. A window in which sympathetic clients were quoted the old rate has tested nothing except your resolve, and its data will flatter the old price.
Expect the mid-test temptation, because it is part of the design: after two refusals, dropping back to $95 will feel like relief. Resist the reading, not just the act. Two answers is weather, not climate - the window was sized precisely because the early results mislead. A test abandoned at the first cold day has cost you the quotes and taught you nothing. The window finishes; then you decide.
4. Reading the result
If enquiries keep saying yes at roughly the pattern the old rate produced, the market has given you the clearest answer it gives: the new number is simply your rate. Promote it - the card already carries it - and let the existing-client rise follow as its own, announced project, in its own time.
If wins collapse, read the reasons before touching the number. A collapse with "too expensive" attached is a different finding from a collapse of timing and silence. Where price really is the objection, work the fix in order: proof first - stronger testimonials and a case study beside the quote give the number something to stand on; then a smaller step - retest at half the step before concluding the old rate was the ceiling; price last, and only on evidence. Most failed tests are thin proof wearing a price tag.
The third result is the mixed one: the new rate wins in one kind of work and loses in another. Believe it. A rate is not obliged to be one number - service lines with different demand and different proof can honestly carry different prices, which is the whole logic of productizing a service into priced packages. The test has not failed; it has drawn you a map.
And do the annual arithmetic before judging any outcome "small". As an illustration, on the experiment's own numbers: if just 300 of your billable hours a year eventually move from $95 to $110, the step is worth 300 x $15 = +$4,500 a year - every year, for a change that cost ten honest quotes to discover. Partial wins are not consolation prizes; priced over a year, they are the point.
5. After the test
Write the decision down, with its date: the rate tested, the window, the results, the number you settled on. It belongs with the rest of your business records, kept the way your documented processes keep everything else - because in a year, when you test again, this page is your control data.
Then end the limbo. The rate card either keeps the new number as the rate or reverts to the old one, deliberately - a card left "in testing" for six months is just an undecided price wearing a lab coat. From that day, the next enquiry hears the settled number with the same calm the test quotes had.
The one courtesy owed is sequencing. If the settled rate is the higher one, existing clients should hear about their own rise from you - with the notice that conversation deserves - rather than discovering the new figure on your public card mid-project. If one of them spots it first and asks, the honest answer is short and sufficient: you have been quoting the new rate to new work while reviewing your pricing, and their own conversation will come with proper notice. It is the experiment's story, told briefly, and there is nothing in it to be embarrassed about.
Questions
Related: the settled rate is an ordinary rate
Once the test settles, the surviving number stops being an experiment and rejoins the ordinary machinery: how to raise your rates handles the existing-client half, your rate card is where the number lives day to day, and the quotes it travels in are built from the proposal template. The experiment's only job was to make those documents honest.