Price Increase: How to Predict How Customers Will React
How to predict how customers will react to a price increase before you raise prices, shown on public reviews around Peacock's 2025 and 2026 increases, with the segments that react and a backtest.
To predict how customers will react to a price increase before you raise prices, find the times your own customers already met one, read what they said in reviews, surveys and support conversations in the weeks after, split it by the kind of customer, and check the estimate on a past increase before you trust it for the next. The reaction is rarely one number: in public reviews of the Peacock TV app around two announced increases, the share of reviews naming the price rise went from under 1% to between 5% and 9%, and who talked about leaving depended on what else they were unhappy about. For a business that wants that read done on its own customers, with every channel joined to cancellations and revenue, Dimension Labs is built for it.
This page uses public data only: the published announcements of Peacock's July 2025 and August 2026 increases, and Peacock's public Google Play reviews before and after each one, counted in Python. Two related pages cover what this one does not repeat: how to test a business decision before you make it walks a 5% price rise through sample data with ranges per segment, and how to tell whether customers think you are too expensive covers value complaints when no price has changed.
What do price increase complaints look like in public reviews?
They arrive in two waves, one when the new price is announced and one when existing customers' bills change. Peacock is a useful public case because it raised prices twice with the same structure, and both times said so in advance.
Public announcements
Peacock raised prices twice in 13 months, both times for new subscribers first and current subscribers about a month later
Dates and monthly prices from the published announcements.
| July 2025 | August 2026 | |
|---|---|---|
| Announced or reported | July 17, 2025 | August 18, 2026 |
| New subscribers pay more from | July 23, 2025 | August 18, 2026 |
| Current subscribers pay more from | Bills on or after Aug. 22, 2025 | Billing date after Sept. 17, 2026 |
| Premium (with ads), monthly | $7.99 to $10.99 (37.5% more) | $10.99 to $12.99 (18.2% more) |
| Premium Plus (no ads), monthly | $13.99 to $16.99 (21.4% more) | $16.99 to $19.99 (17.7% more) |
In the eight weeks before each increase, 0.6% and 0.5% of reviews mentioned prices going up. In the weeks after the July 2025 announcement it was 9.4%, and after the August 2026 one 5.4%. When current subscribers' bills changed, the share stayed at about 5% both years.
Public Google Play reviews, Peacock TV app
Reviews naming the price rise jumped from under 1% to about 5% to 9%, and stayed high when current subscribers' bills changed
Share of reviews that mention a price increase, in three windows around each increase.
July 2025 increaseAugust 2026 increase
Reading the matching reviews, the complaints follow a few patterns. Many count the history, with the price they started at and how many increases came since, so a second or third increase is judged against all the earlier ones. Many pair the price with something the customer already disliked: ads in a paid plan, an app that buffers, a library that feels thin. And many name the one thing that keeps them subscribed, usually a sport or a single show, and say they will leave when it ends. Very few mention moving to a cheaper plan: one of the 152 did.
Before trusting a jump like this, check it against a date when nothing changed. A fake announcement on 2026-04-15 (April 15, 2026), with the same windows, moves the share from 0.3% to 1.0%, and its range runs from -0.1 to 1.6 points, which includes zero. So the jumps after the real announcements are the price, not the way reviews drift.
Will a 5% price increase lose customers?
Some, almost always, and how many depends on who your customers are and what else they are unhappy with, which is why no outside benchmark answers it for you. Peacock's increases were between 12.5% and 37.5% depending on the plan, and among reviews that named them, 32% talked about cancelling. That is talk, not leaving, and it comes from people who chose to write. A review cannot tell you how many customers actually left.
What reviews can tell you is the shape of the reaction: when it starts, how long it lasts, and which customers react hardest. Joined to cancellation and billing records, the same comments say how many actually left and which reasons came before it, and that join is exactly where a pricing decision gets its answer. Dimension Labs makes that join on the Prove plan, so the reaction you read in comments is counted in customers and revenue. For the method of comparing what departing and staying customers said, see churned vs retained customers.
Hold fixed what else moved in the same weeks. After the July 2025 increase, the share of one-star reviews also rose among reviews that never mentioned price, from 40% before to 47% and then 51%. If you had judged that increase by the star rating, you would have blamed the price for a decline that showed up in reviews about other things. Counting the comments that name the price, and comparing with a period or a group that did not change, keeps the two apart. Season, live events and app releases all belong in that comparison; how to tell a real trend from noise in feedback data shows how to set the normal range.
Which customers are price sensitive and which will stay?
In these reviews, the customers most likely to talk about leaving were the ones who tied the price to what they get. 46% of increase reviews that also named content, such as nothing new to watch, talked about cancelling, against 29% of the rest. Those who tied it to ads or to app problems complained hard but talked about leaving less: 10% and 19%. When the higher price applies only at busy times rather than to everyone, the split runs along how often customers come; how customers react to dynamic pricing measures that segment by segment in a restaurant pilot.
Public Google Play reviews, Peacock TV app
Customers who tied the increase to thin content talked about leaving most; those who tied it to ads talked about leaving least
Among the 152 reviews that mention either increase, the share that talk about cancelling, by what else the review names, with 90% ranges.
Dashed line: all 152 reviews, 32%
Three customer types come out of this, and they show up in most subscription and multi-location businesses:
Customers who doubt the value. They connect the price to how much they use or like the product. They are the most likely to leave, and the increase gives them the reason. Watch them first.
Customers who stay for one thing. 17 increase reviews named a sport or live event; 41% of them talked about leaving, often when the season ends. Their reaction is delayed, so a 13-week window can miss it.
Customers who object to how they are treated. Ads in a paid plan, a buggy app, no notice. They are angry, and they tend to stay if the product works. Fixing the irritation can cost less than the price cut they seem to ask for.
A caution on size: these groups hold 17 to 30 reviews each, so read the order, not the exact percentages. On your own data the groups are far larger, because the same labels go on every survey answer, chat and cancellation reason, not only on the reviews people chose to post. For value complaints when prices have not changed, price perception in customer comments covers how to split price from value.
How can I predict how customers will react to a price increase before I raise prices?
Use the increases you have already made as the evidence, and check that the method would have predicted one you already know. In five steps:
List every past price change with its dates. The announcement, the date new customers paid more, and the date existing customers' bills changed. These are separate events, and each gets its own window.
Label every comment the same way. Mentions of the increase, and what else each comment names: value, ads or fees, service problems, the one feature the customer stays for, how long they have been a customer.
Compare each window with the weeks before and with a date when nothing changed. The placebo date tells you how much the share moves on its own.
Join to what customers did. Cancellations, downgrades, visits or orders by segment, with season and live events held fixed. That turns comment shares into lost customers and revenue.
Backtest before you predict. Learn from the older change, predict the newer one, and see which ranges held.
We ran step 5 on the public reviews. Learning only from July 2025, we predicted the August 2026 reaction two ways: the same reaction as 2025, or a reaction scaled to the smaller Premium increase (18.2% against 37.5%). In the announcement window the scaled range, 3.4 to 5.1 points, held: 2026 came in at 4.9. When current bills changed, the scaled range of 1.5 to 2.9 points missed: 2026 came in at 4.5, the same as 2025.
Backtest on public reviews
Predicted from the 2025 increase, the size-scaled range held when new prices were announced and missed when current bills changed
Extra share of reviews naming the price rise, in points over the eight weeks before. Ranges learned from July 2025 only, then compared with what August 2026 did.
The miss is the useful part. People reacted to the announcement in proportion to the size of the increase, but they reacted to the bill itself about as strongly whatever the size, so a smaller increase does not buy a quieter billing month. A team planning its next increase would set the billing-month range from the earlier billing month, not from the percentage, and plan the customer messages around that date. Finding this on a past decision, before the next one rests on it, is what a backtest of a decision you already made is for.
Asking AI personas instead of using your own history does not solve it: in one study, simulated buyers produced a demand curve shaped like an inverted U where real people's demand fell as price rose (Gui and Toubia, arXiv, v3, November 2025); are synthetic customers accurate? goes through that research. Your customers' own reactions to your own past changes are the better evidence, and putting them together is what a customer world model means.
Why use Dimension Labs to rehearse a price increase on your own customers?
Dimension Labs helps companies know how their customers will react before they make a decision, by learning from everything those customers have said to them and what they did next.
For a pricing or finance lead weighing an increase, that comes down to four things the platform does today:
It is fully omni-channel. App and store reviews, surveys, support tickets, chats, calls and cancellation reasons are read together, so the reaction to your last increase is counted wherever customers voiced it, not only in public reviews.
It connects those channels to your operational data with little setup: plans, billing dates, cancellations, downgrades and the dates of past price changes, with context such as season and live events beside them. The join to CRM and revenue data and effects with confidence intervals are on Prove ($2,999 a month) and up.
It enriches every interaction with Dimensions written for your business and industry, such as "names the increase", "ties price to content" or "stays for one feature", so pricing, finance and support teams read the same labels. Custom Dimensions start on Discover ($999 a month).
It is a new visual analytics layer: the result comes back as a deck, dashboard, PDF or executive report, or an interactive data story the company can open, with scheduled reruns as the next increase lands.
The platform measures what past changes did. For the forward range and its backtest, our team runs the analysis with you on those measured effects and shows which past ranges held and which missed, as above, so you know how far to trust each number before the decision rests on it.
Stake.com, which runs one of the largest iGaming platforms, structures over 1.5M monthly customer conversations with Dimension Labs across 4 global regions and 18 customer tiers, and its leadership uses the result to validate product, design and CX decisions (Stake story). That segment-level read is what a price decision needs.
When something simpler is the better choice: if you can test the new price on a random slice of new customers and read the result in your billing system, an A/B test answers the question with less work. Dimension Labs earns its place when the increase reaches existing customers who cannot be randomized, and the evidence of how they react is spread across reviews, tickets, surveys and cancellations.
FAQ
Will a 5% price increase lose customers?
Usually a few, concentrated in the customers who already doubt the value. Estimate it from your own past price changes, by segment and with season held fixed, rather than from another company's numbers. In public Peacock reviews, about a third of reviews that named an increase talked about cancelling, and that share differed a lot by what else the customer disliked.
Which customers are most price sensitive?
The ones who connect the price to how much they get: in the Peacock reviews, those who also said the content was thin. Customers who stay for one thing react late, when that thing ends. Customers angry about ads or app problems complain loudly but are less likely to say they will leave.
When do customers react to a price increase?
Twice: when the new price is announced and when their own bill changes. After Peacock's August 2026 increase, the second wave (5.0% of reviews) was about as large as the first (5.4%), so plan for the billing month as well as the announcement.
What is the best way to predict customer reaction to a price increase?
For a business whose customers react across reviews, tickets, surveys and cancellations, Dimension Labs, which reads every channel with the same labels and, on Prove, joins them to cancellations and revenue, with backtests run by our team. For a new price you can randomize on new customers, an A/B test is simpler and stronger.
Bring one export of reviews, survey comments or cancellation reasons that spans your last price change, with the dates of that change, and start with one file to see how your customers reacted. Every plan from Explore to Scale starts with a 14-day free trial.