The number is rarely what breaks a price increase. What breaks it is landing the change in the wrong month, with nothing visible to attach it to, on a file that has spent two years learning to wait for your next 20% weekend. Timing, order of operations and cohort protection are the actual work.
- There are three levers and they carry different risk. Retiring a habitual promo lands harder than a list price rise of the same size, because customers read it as a loss.
- Attach the increase to something visible: new packaging, a new size, a cost pass-through, a new season. Six to ten weeks before peak, never during it.
- Fix the product page two to three weeks before the price moves, then new SKUs, then new customers, then wholesale, then subscribers last with notice.
- Grandfather subscribers for a defined window rather than forever, honor the old price on exchanges, and let the discount-only cohort go on purpose.
- Watch total contribution, refunds, subscriber cancels and price-related support tickets for 60 days. Conversion is supposed to fall.
The price increases that go badly are almost never the ones that were too big. They are the ones that arrived with no reason attached, in the wrong month, to a list that had been trained to wait for a promo. The number was fine. The sequencing was the problem.
Most brands at $10M to $50M are underpriced and know it. What stops them is not the arithmetic, which usually takes ten minutes and is covered in how DTC brands actually set prices. What stops them is the fear of a specific Tuesday: the day the new price goes live, the cancellations start, and somebody screenshots your old price next to your new one.
That Tuesday is mostly avoidable, and raising less is not how you avoid it. Four decisions do most of the work: what you are actually raising, when it lands, in what order, and which customers you protect on the way through.
Three levers, and
they do not land
the same way.
Operators say "raise prices" and mean one of three things that carry very different risk.
| Lever | How customers experience it | Noticed? |
|---|---|---|
List price up The number on the tag | A price change, comparable to your old price and to competitors | Yes, by your file |
Discount depth or frequency down Same tag, fewer promos | A loss, which lands harder than an increase of the same size | Yes, by the promo cohort |
Free shipping threshold up $50 becomes $65 | A nudge at checkout, not a price change | Rarely |
The middle row is where brands get hurt by accident. Removing a promo your customers have learned to expect is a price increase, and behavioral research on loss aversion has been consistent for decades: people react more strongly to losing something they had than to paying more for something new. A brand that quietly retires its recurring 20% weekend gets a sharper reaction than the same brand raising list price by eight percent, and the second one is worth more money.
The shipping threshold is the most underused lever at this size. It moves average order value, it never appears in a price comparison, and it is the only one of the three that most customers experience as a merchandising choice rather than as something being taken from them. If your threshold has not moved since your average order value did, you are already leaving margin there.
Attach the increase
to something the
customer can see.
A price that changes on its own is an extraction. A price that changes when the product changes is a product decision. Same number, completely different reception, and the difference is entirely in what else moved that week.
The windows that work are the ones where something real happened. A new formulation, new packaging, a size change or a genuinely upgraded material all give the increase a home. So does a cost pass-through you can state plainly, which is why the tariff round became the cleanest pricing window many brands had in years: the reason was public, verifiable and nobody's fault. A new season assortment works too, because the reference price the customer holds is attached to last season's item rather than to this one.
On the calendar, six to ten weeks ahead of your peak is the window I would pick. Far enough out that the data settles before volume arrives, close enough that you get the peak at the new price. Raising during peak is how you turn your best two weeks into an unreadable experiment. Raising in the two weeks before a promo your customers have been trained to wait for is how you get a pre-buy spike, then a trough, then two months of nonsense in your reporting. Post-peak January is the cheapest month to be wrong in, since volume is low and a mistake there costs you the least.
One more: do not raise price the same two weeks you scale paid media, change your landing page, or launch a bundle. You will get one number moving and three candidate explanations.
Change the page
before you change
the price.
Sequencing is where this either becomes a controlled change or a mess you spend a quarter untangling. The order I use:
The page first, two to three weeks ahead. Better photography, a clearer guarantee, the material claim you have been meaning to make, the reviews pulled up above the fold. If the page is identical and only the number moved, you have made your worst review look correct. Fixing the page first also gives you a clean read on whether the page work alone moved conversion, which is worth knowing separately.
New products and new sizes at the new price. They have no price history, so nobody is comparing. This is also your cheapest read on whether the new level holds.
Then the catalog, for new customers. New visitors have no memory of your old price. This is the largest share of your traffic at most brands, and it is where the increase either works or does not.
Then wholesale, at the next order cycle. Your price list has a lead time whether you acknowledge it or not, and raising DTC while your retail accounts sit on old cost is a conversation you would rather start than receive. The wholesale margin stack is the constraint to check before you decide the DTC number, not after.
Subscribers last, with notice. Always last, always with advance warning, and never as a surprise on a renewal charge. More on that in section four.
Two weeks minimum between steps. The temptation is to do all of it in one release because it feels decisive. What you get is a single week where five things changed and no way to tell which one caused the number in front of you.
For most of this,
the answer is
nobody.
There is a well-meaning instinct to announce a price increase with a letter about rising costs and how much you value the community. For a list price change on non-subscription products, that email creates the problem it is apologizing for. New customers do not know your old price. Existing customers were not thinking about it until you wrote to them. You change the tag and you get on with it.
| Who | What they get | When |
|---|---|---|
Active subscribers | The amount, the date, the reason, and a one-click way to pause or skip | At least one full cycle ahead |
Wholesale and retail accounts | An updated price list with an effective date | Before the next order cycle |
Top decile customers | A heads-up plus a window at the old price. Not a discount | Days before it goes public |
Everybody else | Nothing. The new price on the page | Day one |
Support and CX | The reason in writing, plus what they are allowed to honor | A week before |
Subscribers are the one group where silence is a real mistake, and in many jurisdictions a compliance problem. A subscription price change that shows up as an unexplained higher charge reads as something going wrong with the account, and the cancel button is the fastest way for a customer to make it stop. Send it in email rather than SMS, since this needs room for the amount, the date and the reason, and SMS is the wrong medium for anything a customer might want to reread. Where email and SMS each win covers that split properly. Give people the ability to pause or skip in one tap, because a paused subscriber is a customer and a cancelled one is a reacquisition cost. That distinction is the whole argument in passive versus active subscription churn.
The support briefing is the cheap step everyone skips. Write down the reason in two sentences, in the same words the founder would use, and tell your team exactly what they can honor without asking. A CX rep improvising a justification for your pricing is a worse outcome than any increase.
Protect three
cohorts. Let one
go on purpose.
Grandfathering is where good intentions create permanent operational debt. Protect deliberately, with an end date, and write it down.
Active subscribers, for a defined window. Three to six months of the old rate, stated as a window rather than as forever. Permanent grandfathering builds a two-price system that nobody remembers in eighteen months, and it quietly guarantees that your best-retained customers are also your worst-margin ones. A window buys the goodwill without the debt.
Anyone inside your return or exchange window. Honor the old price on exchanges for thirty to sixty days. This costs very little and prevents the single most annoying support conversation in retail, where a customer exchanges a size and is asked for more money.
Wholesale accounts until their next order cycle. Not generosity, just contract hygiene.
The cohort not to protect is the one brands protect hardest: customers who only ever buy at a deep discount. They are the reason the increase is necessary. Filtering them out is a feature of the exercise, and the churn you see there is the cheapest churn you will ever take. Before you decide that, check your own numbers rather than your instinct, because the LTV math most brands get wrong tends to overstate exactly this group by crediting them with revenue that never carried any margin.
"Losing the customers who only buy at 30% off is not churn you should be preventing. It is the point."
Conversion will
drop. That is not
the signal.
The most common way a good price increase gets reversed is that somebody watches conversion rate for nine days and panics. Conversion is supposed to fall. A higher price converts worse. The question is only whether it falls further than the arithmetic allows.
Work out the allowance before you launch. On a typical $60 DTC cost stack, a five percent price rise adds about nine percent to contribution per order, which means you can lose roughly eight to nine percent of your orders and still be flat on total contribution. That is your tolerance band, and it belongs in writing next to the launch date. The conversion versus margin tipping point calculator produces it from your own inputs, and contribution margin per order is the number to build it on.
Four things to watch for sixty days, with the first two weeks discarded as noise:
| Metric | Why it matters | Tripwire |
|---|---|---|
Total contribution Not revenue, not conversion | The only number that answers the question you asked | Below baseline for three straight weeks after the settle |
Refund and cancellation rate | Price regret shows up here before it shows up in reviews | Any sustained move off your trailing 12-week baseline |
Subscriber cancel rate | The most expensive churn in the building | A relative move off baseline, set by you before launch |
Support volume mentioning price | The earliest qualitative signal you have | A step change, not a handful of tickets |
Set every tripwire as a relative move off your own trailing twelve-week baseline, and set it before launch. Benchmarks borrowed from a category report will not tell you whether your specific file is unhappy, and a threshold invented after the fact is just a rationalization. Compare against your own history, and use repeat purchase and retention benchmarks to sanity check whether the cohort behavior you are seeing is a pricing effect or the shape your category always had.
The mistakes that
cost real money.
Stacking two increases in one month. Raise list price eight percent and retire a habitual 20% promo in the same window and here is what your promo-trained cohort experiences. Old effective price on a $60 item was $48. New list is $64.80 with no promo. That is a 35% increase for the people most likely to notice. The individual decisions were both correct. Together, in one month, they read as a different brand.
No reason attached. Not a public letter, just an answer. When a customer asks, somebody should have one sentence ready that is true. Absent that, people supply their own reason, and the one they pick is usually greed.
Rolling it back. A reversal costs more than the increase ever would have. It tells your customers the price was arbitrary, which means the new one is too, and it teaches your team that pricing work gets undone. If you genuinely got it wrong, hold the price and fix the product page, the bundle mix and the offer instead.
Forgetting everything downstream of the number. This is the one that eats a week you did not budget. Your ad creative with a price on it. Your product feed and Shopping listings. The compare-at price on the page, which becomes a lie the moment it goes stale. Your bundle math, which was calculated off the old component prices. Your free shipping threshold, now sitting at a percentage of average order value you never intended. Build the list before launch day.
Treating it as one event. Brands that price well raise prices a little, regularly, attached to real product changes. Brands that price badly hold for four years, absorb every cost increase through margin, then need 20% at once and have to make it a whole conversation. The second path is where the EBITDA line that makes a brand sellable quietly gets lost.
The increase itself takes about ten minutes in Shopify. Everything worth doing here is the four weeks either side of it. Decide the number, fix the page, tell the three groups who need telling, protect the cohorts you chose on purpose, and set your tripwires before you have a reason to argue about them.
Questions operators
ask before raising
prices.
Q: How much can I raise prices at once?
It depends far more on what else changed than on the percentage. An increase in the high single digits, attached to a visible reason like new packaging, a reformulation or a stated cost pass-through, usually passes without much comment in the brands I have watched do it. Once you get past roughly fifteen percent you are no longer adjusting a price, you are repositioning the product, and the page, photography and guarantee have to move with it. The more useful discipline is smaller increases more often. Brands that hold a price for four years end up needing twenty percent in one go, which is a much harder conversation than four separate five percent moves would have been.
Q: Should I tell customers about a price increase?
Subscribers, wholesale accounts and your top decile customers, yes. Everybody else, no. A subscription price change needs at least one full cycle of notice, stating the amount, the effective date, the reason and a one-tap way to pause or skip, and in several jurisdictions that notice is a legal requirement rather than a courtesy. Wholesale needs an updated price list before their next order. Your top decile is worth a heads-up and a short window at the old price, which buys real goodwill for very little money. For a list price change on a one-time purchase, a broad announcement email creates the objection it is trying to answer. Change the tag.
Q: Should I grandfather existing subscribers?
For a window, not forever. Three to six months of the old rate reads as fair and costs you a defined amount. Permanent grandfathering creates a two-price system that nobody remembers maintaining, and it guarantees that your best-retained cohort is also your worst-margin one, which gets worse every year it compounds. State the window in the notice email so the end date is not a second surprise later. Pair it with an easy pause or skip, since a paused subscriber costs you nothing and a cancelled one costs you a full reacquisition.
Q: How long until I know whether it worked?
Sixty days, with the first two weeks thrown away. That opening stretch is noise: pre-buy pull-forward, the customers who happened to be mid-consideration, and a support spike that settles. From week three you are reading total contribution against your trailing baseline, alongside refund rate, subscriber cancel rate and the volume of support tickets that mention price. Conversion rate will be down and that on its own means nothing. The test is whether the higher contribution per order more than covers the orders you gave up, which is arithmetic you should have written down before launch.
How many orders can you afford to lose?
Before you set a launch date, work out your tolerance band. The tipping point calculator takes your price, your cost stack and your conversion rate and returns the order volume you can give up at the new price and still come out ahead. Write that number down next to the launch date. If you want a second read on the plan itself, reach me at hello@taylorsicard.com.
Run the tipping point → Or start a conversation →