We diff vendor pricing pages against their last verified state, every cycle, for 482+ products. In 2026 that produced 3,843 logged changes. The headline is not the hikes. It is that vendors rebuild their tier structure about 4 times more often than they touch a price.
Changes logged
3,843
2026, 482 products
Plans added / removed
1,831 / 1,163
packaging churn
Price moves
749
hikes lead 178:80
Median hike
+23%
vs -21% median cut
If you only remember one number from this report, take 4. For every list-price change we logged, vendors made about 4 changes to what the plans contain. A price hike is a press release and a wave of "reconsidering our subscription" tickets. Retiring a tier, splitting a feature into an add-on, or inserting a new plan above the popular one achieves the same revenue result with none of the noise. The pricing page keeps its numbers; the invoice grows anyway.
The direct hikes that do happen are mostly small and unannounced. The median move in our plausible band is +23%, the kind of step that slips under a finance review. The exceptions are worth watching precisely because they are rare: Figma taking Professional from $15 to $20 is a third more revenue from the exact plan most paying teams sit on. On the other side, Jira cut Standard and Premium by double digits in the same month. When a market leader cuts seat prices, someone in that market is winning on price.
A word on honesty, because volatility data is easy to inflate. Some of the biggest deltas in our raw log are not market moves at all: they are our own tracker re-baselining a promo rate to the list rate, or a vendor rebuilding a broken page. Whereby Business "moving" from $11.99 to $60 and back inside ten days is a page in mid-surgery, not a 400% hike. Headline statistics on this page therefore exclude moves beyond 50% either way. The raw events stay in the public changelogs, labelled with dates, so you can judge them yourself.
What should a buyer do with this? Screenshot nothing, bookmark the changelog. Any quote you hold is a snapshot of a moving object, and the object moves monthly. If a vendor reshuffles tiers twice a quarter, expect your renewal to arrive with a new plan name and a "grandfathering" clause worth reading twice.
Went from $15 to $20 per seat on June 28, a 33% hike on the mid tier, weeks after the same plan briefly listed at $12. Two moves in one quarter on the plan most teams actually buy.
Moved the other way: Standard from $9.05 to $7.91 and Premium from $18.30 to $14.54 in late June. Seat-price cuts of 13-21% are what defending market share against flat-priced rivals looks like.
Listed at $11.99, then $60, then $13.99 within ten days in late May. That is not a strategy, that is a pricing page being rebuilt in production. We record what the page said on each date.
Plan-level events (price changes, additions, removals) by category, 2026 year to date.
| Category | Plan events in 2026 |
|---|---|
| Cloud Hosting | 232 |
| Large Language Models | 204 |
| Website Builders | 169 |
| Design Tools | 166 |
| AI Image Generators | 156 |
| Databases | 127 |
| Email Marketing | 117 |
| AI Productivity | 115 |
| CRM | 114 |
| Project Management | 111 |
| Accounting | 111 |
| Video Conferencing | 108 |
Constantly, but not where you expect. Across 2026 we logged 749 list-price changes against 2,994 packaging changes (plans added or removed). The tier structure churns roughly 4 times more often than the sticker price.
Among plausible list-price moves we tracked, hikes outnumber cuts 178 to 80, roughly 2.2 to 1. The median hike is 23% and the median cut is 21%. Prices drift up, quietly and in single-digit steps, punctuated by the occasional 30% jump like Figma Professional.
A price change is news and invites churn reviews. A packaging change is invisible: retire a tier, split a feature into an add-on, introduce a new "recommended" plan one notch higher. The invoice grows without the per-seat rate ever moving, which is why plan adds and removals outnumber price moves in our data.
We re-verify vendor pricing pages on a rolling cycle and diff every plan against the previous verified state. Each change lands in a product changelog with the date, the plan and both values. The histories are public on each product page.
Every verification cycle diffs each product's plans against the previous verified state; differences are logged as dated changelog events with before and after values. Headline hike and cut statistics use only moves within 50% either way: larger deltas are dominated by promo-to-list re-baselining and page rebuilds rather than market repricing, and are excluded from summary figures while remaining in the public per-product changelogs. All numbers compute from the changelog at build time, as of July 2, 2026. Full methodology.