How to raise subscription prices without a churn spike#

Two facts decide most of this. WooCommerce Subscriptions grandfathers your existing book automatically - Woo's own developer FAQ states that if you change a subscription product's price, "All existing subscriptions will continue to be charged the old price by default." And if you decide to migrate them anyway, California requires notice of a fee change no less than 7 and no more than 30 days beforehand, while New York requires it at least 5 business days and no more than 30 days ahead. Everything else is sequencing.

Price increases fail in a predictable way. The merchant announces on a Tuesday, the increase lands on the next renewal, and the cancel spike shows up two to six weeks later spread across renewal dates, by which point nobody can attribute it. This post is the order of operations that avoids that, plus an honest read on what the public data can and cannot tell you about the size of the hit.

What the public data actually supports#

There is no published benchmark that says "a 10% price increase produces X% incremental churn" for ecommerce subscriptions. Figures phrased that way circulate widely and trace back to nothing - I chased several and found no primary study behind any of them.

The best public evidence is a measured time series rather than a benchmark. Antenna, which estimates US subscription behaviour from a consumer panel, published Netflix's monthly churn around two price increases in A Whole New Netflix (accessed August 2026). Netflix stopped reporting quarterly subscriber counts, so these are Antenna's panel estimates, not company-disclosed numbers.

The shape is the finding. A price increase produces a churn spike in the month it lands, and the spike decays over roughly the following four months without returning all the way to baseline. The narrow 2023 increase moved churn 0.2 points. The broad 2025 increase, which touched every US plan, moved it 0.7 points and took four months to fall back near the prior level. Breadth of exposure mattered more than the size of the increase.

Do not read Netflix's absolute churn levels as a target for your store. Different category, different price point, different substitutes. Read the shape.

One more finding worth holding: McKinsey's February 2018 subscription survey of 5,093 consumers found "value" was the most-cited cancellation reason for curation and access subscriptions, at 29% each. Price increases do not attack convenience or habit. They attack the value judgment, which was already the largest single reason people leave. That is why a price increase paired with no visible change in what the subscriber gets is the version that hurts.

Grandfathering is the WooCommerce default, whether you chose it or not#

On WooCommerce Subscriptions, the recurring amount is calculated and stored on the subscription at the time of purchase. Edit the product price and nothing happens to live subscriptions. The WooCommerce developer FAQ says it plainly: existing subscribers are not affected, and all existing subscriptions continue at the old price by default. Applying a new price to the existing book means editing subscriptions individually or running a bulk update through automation tooling.

This is a good default and a dangerous one.

Good, because it means you can raise your list price today with zero churn risk. New customers pay more, nobody who already trusts you gets a surprise, and your blended ARPU rises as the book turns over.

Dangerous, because grandfathering compounds silently. Three years of unmigrated increases produces a long tail of subscribers on prices you would never sell today, and they are usually your most loyal cohort - the exact people you least want to shock when you finally do migrate them. Every year you defer, the eventual migration gets bigger and the gap you have to close in one step gets wider.

The two strategies, honestly compared#

Grandfathering trades revenue for zero churn risk; migration does the reverse. No credible public benchmark exists comparing the two - nobody has published a controlled study of grandfathered versus migrated cohorts in subscription commerce, so the table below reasons from mechanics rather than from measured outcomes. Treat it as a structure for your own decision, not as evidence.

The middle path most stores should take: grandfather by default, migrate in cohorts with an announced cadence, and never migrate more than one price step at a time.

The notice rules are not optional#

If you migrate existing subscribers, you are changing the terms of a running auto-renewal, and several states regulate that directly. Two you cannot ignore:

California. Business and Professions Code section 17602 requires that when there is a change in the fee, the business provide, "no less than 7 days and no more than 30 days before the fee change takes effect," a clear and conspicuous notice of the fee change, along with information on how to cancel. The same section requires a separate clear and conspicuous notice of any material change to the terms.

New York. General Business Law 527-a requires clear and conspicuous notice of "any material change to the terms of the automatic renewal or continuous service offer, including any price increases," delivered "at least five business days prior, but no more than thirty days prior, to the date of the change," in the same medium you used for renewal notices. New York also pairs this with a cancel-and-pro-rata-refund backstop.

Note the shape of both windows. There is a floor and a ceiling. Sending the notice six weeks out is not extra-safe in California or New York - it is outside the window, and you would need to send again inside it. Plan the send date, not just the send.

There is no empirical research on whether a longer notice period reduces cancellation. It is an obvious question and nobody has published an answer, so pick your send date from the statutes rather than from a retention theory.

The wider state map, including which states have renewal-reminder obligations that interact with this, is in state auto-renewal laws. This post assumes you have read it.

Sequencing: the order that works#

The increase should be the last thing that happens, not the first.

1. Ship something first. Two to eight weeks before the notice, land a visible improvement - a better box, a new destination, faster support hours. The improvement does not have to justify the increase economically. It has to exist, so the notice email has something to point at other than costs.
2. Raise the list price for new customers only. Let it run for at least one full billing cycle. This gives you a clean read on whether the new price hurts acquisition before you also expose the existing book to it.
3. Pick the cohort. Oldest first, on the widest price gap. Migrating your newest subscribers first is backwards: they have the least accumulated goodwill and the smallest gap.
4. Send the notice inside the statutory window. One email, subject line that says the price is changing, the old price, the new price, the date of the first charge at the new price, and a cancel link. Do not bury it in a newsletter.
5. Send a reminder 3 to 5 days before the first new-price charge. This is the one most stores skip. The renewal charge should never be the first time a customer remembers the notice.
6. Hold the cancel flow steady. Do not launch a new save offer in the same week. You will not be able to tell which change moved the number.
7. Measure at renewal date, not announcement date. See the measurement section below.

Framing: what to say and what not to say#

Three rules that fall out of the value-judgment framing:

• Name the number in the subject line. "Your subscription price is changing on 12 October" outperforms cheerful evasion because the customer who was going to be annoyed is going to be annoyed either way, and the ones who find out at the charge are the ones who cancel and dispute.
• Do not lead with your costs. Carrier rates, ingredient costs and hosting bills are your problem, not the subscriber's. One clause is fine. A paragraph reads as pleading and invites the reply "then absorb it".
• Give an off-ramp that is not cancellation. Annual prepay at the old rate, a smaller plan, or a pause. If a subscriber's only two options are the higher price or the exit, some meaningful share will take the exit purely because it was offered as a binary.

The second rule has some experimental support, from an unexpected industry. A 15-cell field experiment run with 1,655 customers of a multi-site Canadian self-storage provider tested cost-based, quality-based and market-based justifications for a price increase. Contrary to the usual advice, market justifications produced the lowest attrition. The published abstract does not give effect sizes, so the finding is directional only, and self-storage is not subscription commerce - customers there have switching costs a box subscriber does not. Take it as one piece of evidence against the reflex to explain your cost problem, not as a rule.

The annual-prepay off-ramp is the strongest of the three and the most often forgotten. It converts a price objection into a term commitment, which is a trade most price-sensitive subscribers will take. The retention arithmetic behind that swap is in annual plans and churn - including why the measured churn improvement is partly mechanical, which matters if you plan to grade your own price increase on the churn number afterwards.

What to do with the cancel flow#

A price increase produces a burst of cancel-intent traffic with an unusually homogeneous reason: too expensive. Two adjustments for the migration window:

• Add a temporary reason option that names the increase directly, separate from the generic "too expensive". Otherwise the increase's cancels are indistinguishable from your baseline price cancels and you will never size the effect.
• Route that reason to a term offer, not a discount. Discounting the customer straight back to the old price on request converts your price increase into a price increase for people who did not complain. That is the worst of both worlds and it teaches the behaviour.

What to measure#

The measurement mistake is comparing the month of the announcement against the month before. Renewal dates are spread across the month, so the effect smears. Measure by renewal cohort instead.

• Cancel rate per renewal cycle, for subscribers on the new price vs still on the old price, over the same calendar window. This is the only comparison that isolates the increase.
• Net revenue at cycle 1 and cycle 3. A 12% increase that costs 5% of the cohort is still up. A 12% increase that costs 15% is not. Cycle 3 matters because delayed cancellation is common: some subscribers take one charge at the new price before deciding.
• Involuntary churn in the migration window. A higher charge amount fails against limits and stale cards more often than the old one did. Watch decline rates, not just voluntary cancels.
• Support ticket volume per 1,000 notified. If this spikes, your notice email was unclear, and the fix is the email rather than the price.

If you are also moving shipping rates in the same period, do not do both at once - shipping costs are their own churn lever, and stacking them makes the post-mortem impossible.