# Nine Changes, Not One a Price: Re-Reading Nine WordPress Pricing Pages Weekly in 2026

> Nine WordPress performance pricing pages, re-read on seven dated occasions from 26 September to 2 November 2026. Ten changes, and not one was a list price.

- Published: 2026-11-02
- Updated: 2026-11-02
- Author: M Asif Rahman (https://ar.bd), Founder, WPDeveloper
- Tags: Pricing, Research, Methodology, Market Data, Comparison, WordPress, platform:wordpress
- Canonical: https://xspeedcache.com/blog/vendor-term-volatility/

---

Updated November 2026

Between 26 September and 2 November 2026 we re-read the same nine WordPress performance pricing pages on seven dated occasions and recorded ten changes. Not one was a change to a published list price. Five of the nine pages recorded no change in any field we track, and one vendor accounted for six of the ten.

The comparison every buyer expects is which caching plugin costs least. The more useful question, which no comparison answers, is how long any price on any of those pages stays true. This post is about the pages rather than the products, and every figure below is cited to the vendor's own page with the date it was read.

## Quick Summary: Which Numbers to Trust, and for How Long

| If you want to… | Do this | Why |
|:---|:---|:---|
| Budget a licence this month | Read the list price off the vendor's page | List prices were the most stable field: zero moved |
| Know what you will pay today | Check for a promotion, a code and a multi-year toggle | Two vendors moved their effective price without touching a list price |
| Know what year two costs | Read the renewal rule, a separate sentence from the price | One vendor's FAQ limits the discount to a first purchase |
| Quote a price in writing | Quote it with the date you read it | WP Rocket's headline price expires 3 November 2026 |
| Trust a price series | Ask how many readings it has, and whether it publishes errors | Two readings here were wrong, and both errors were ours |

The short version: the number is the stable part. The conditions attached to it are what move, and a comparison table almost never carries those.

## How We Read the Same Nine Pages Seven Times

The method is deliberately dull, because reproducibility is the point. Nine pages are fetched with `curl` under a desktop Chrome user agent, converted to text, whitespace-normalised, then read by hand for five fields and written into this engine's notes with the date.

The nine, each linked to the page actually read: [WP Rocket](https://wp-rocket.me/pricing/), [WP Fastest Cache](https://www.wpfastestcache.com/), [WP-Optimize Premium](https://teamupdraft.com/wp-optimize/pricing/), [W3 Total Cache Pro](https://www.boldgrid.com/w3-total-cache/), [FlyingPress](https://flyingpress.com/pricing/), [Perfmatters](https://perfmatters.io/pricing/), [NitroPack](https://nitropack.io/pricing/), [Object Cache Pro](https://objectcache.pro/pricing) and [Hummingbird Pro](https://wpmudev.com/pricing/). Three need handling a scraper will not guess: WP Fastest Cache keeps its figures on its homepage, and `objectcache.pro` answers 403 to `curl`.

The five fields, recorded per vendor per reading:

- 🏷️ **List price** per tier, with the licence unit in the vendor's own words.
- 🏷️ **Effective price today**: promotion, code, countdown or multi-year prepay, and its stated end date.
- 🚪 **Entry condition**: whether a trial exists, and whether it charges automatically.
- 🚪 **Exit condition**: the refund window, guarantee or remedy, and what a lapse does.
- 🔁 **Renewal rule**: whether the price paid repeats, and any published renewal discount.

The readings fall on six consecutive Saturdays (26 September, then 3, 10, 17, 24 and 31 October) plus a seventh on Monday 2 November 2026, taken for this article. Two earlier partial readings, on 7 and 9 September, are a baseline, not part of the series.

## What the Instrument Cannot See, Stated Before the Results

**A change can only be dated to the gap between two readings.** Where a term moved between 17 and 24 October, the series cannot say which day. Eight days is the resolution.

**Nine vendors are not the field.** These are the pages this engine re-reads for its own comparisons, so it is a convenience sample and nothing below generalises to the whole market.

**A page can read differently to a script than to a human.** WP Rocket's promotional page renders `$ 5 / month` and `$ 2 .97`, so a match run without whitespace normalisation reads prices that do not exist. Its Multi card carries a 50/100/500-site selector, which is why one reading recorded five rungs and another three. Neither was wrong.

**A cached copy can serve a figure the vendor has already replaced.** Pricing pages are ordinary HTTP resources under the freshness rules [MDN's HTTP caching guide](https://developer.mozilla.org/en-US/docs/Web/HTTP/Guides/Caching) sets out, so a reading is evidence about one fetch at one moment.

**The series records the instrument as well as the subject.** Two readings were wrong, both errors ours, and one was found while this article was being written.

## The Term-Volatility Matrix: Nine Pages, Seven Readings, Ten Changes

Every difference between consecutive readings, oldest first. `P` is a presentation change, `T` a term change, `E` an effective-price change.

| # | First seen | Vendor | What changed | Class |
|:---|:---|:---|:---|:---:|
| 1 | 3 Oct | WP Rocket | The 100- and 500-site prices became readable at $399.95 and $599.95; on 26 September those tiers carried no figures | P |
| 2 | 3 Oct | WP-Optimize | The unlimited tier is labelled "Enterprise"; $194.00 did not move | P |
| 3 | 10 Oct | WP Rocket | *"We don't offer a trial version"* became *"Yes, we now offer a free trial"*: 14 days, charging automatically unless cancelled | T |
| 4 | 17 Oct | Perfmatters | A public code appeared above the ladder, worth an extra 10% off | E |
| 5 | 24 Oct | WP Rocket | $59.95 / $119.95 / $299.95 a year became $5 / $10 / $25 a month *"Billed yearly"*, annualising to $60 / $120 / $300 | P |
| 6 | 24 Oct | WP Rocket | A three-year prepay at 30% off, $3.49 / $6.99 / $17.49 a month, billed upfront | T |
| 7 | 24 Oct | FlyingPress | *"No refunds needed"* replaced by a scale: 100% within 24 hours, 80% up to 7 days | T |
| 8 | 24 Oct | FlyingPress | A renewal-notice term appeared: an email reminder 30 days before renewal | T |
| 9 | 31 Oct | WP Rocket | A countdown promotion: up to 40% off, one year at $4 / $8 / $20, three years at $2.97 / $5.94 / $14.87 | E |
| 10 | 31 Oct | FlyingPress | The refund scale completed with its third rung: *"none after 7 days"* | T |

![Grid of nine vendor pricing pages against seven dated readings, marking which reading first recorded each of the ten changes.](https://xspeedcache.com/images/blog/vendor-term-volatility-matrix.webp)

Five of nine pages contributed nothing, and stability is a finding too. Through all seven readings WP Fastest Cache held $49 / $125 / $175 / $300 one-time, W3 Total Cache Pro held $99 / $350 / $500, NitroPack held Free, $8.00, $22.00 and $99.99 a month, Object Cache Pro held $95 a month, and Hummingbird Pro published no standalone price at all.

## Five Classes of Change, and Only One of Them Is a New Price

That matrix is readable only because every row is classified. A tracker that records "the price changed" and nothing else reports changes that did not happen, and misses the ones that did.

- 🏷️ **Price move.** A list figure changed. None in six weeks, beyond a five-cent rounding.
- 🏷️ **Effective-price move.** The list figure held; what you pay changed. Perfmatters' code, 17 Oct; WP Rocket's promotion, 31 Oct.
- 🚪 **Term move.** Both prices held; a condition changed. FlyingPress replacing "no refunds needed" with a scale.
- 🔁 **Presentation move.** Nothing changed; the page reads differently. WP Rocket restyling a yearly ladder as a monthly one.
- 📌 **Reading error.** The page did not change; our earlier reading was wrong. Two of ours, both below.

![Bar chart of the ten changes by class: five term, three presentation, two effective-price and zero list-price moves.](https://xspeedcache.com/images/blog/vendor-term-volatility-classes.webp)

Sorted by class: two effective-price moves, five term moves, three presentation moves, zero price moves. **The published number was the most durable thing on these pages, and the conditions attached to it were the least.**

## WP Rocket Moved Six Things Without Moving Its Ladder

One vendor accounts for six of the ten rows, and reading them in order shows what a six-week window buys.

On 3 October the page read, verbatim: *"We don't offer a trial version, but we do have a 14 day refund policy."* By 10 October the same question answered *"Yes, we now offer a free trial"*, the vendor's own word "now" doing the dating. That falsified a live post of ours inside a week: we had published that a competitor was the only product here you could try before paying, and a refresh on 9 October cut the claim. By 24 October the ladder was restyled from yearly figures to monthly ones, with a three-year prepay at 30% off. By 31 October that had deepened to 40%, acquired a countdown, and acquired a sentence that matters more than any price on the page:

> "The Halloween discount is applied when you buy WP Rocket for the 1st time. Once your initial license period ends, your license will renew annually at the regular renewal price."

Read on 2 November 2026, the page still carries a banner over a running timer:

> "Up to 40% OFF. Hurry up, the offer is only valid until November 3, 2026!"

That is a vendor publishing the expiry date of its own headline price. **Any comparison quoting this vendor at $2.97 a month becomes wrong tomorrow.** Our own [three-year cost model across nine caching licences](https://xspeedcache.com/blog/wordpress-caching-plugin-cost-comparison/) prices the renewal rule rather than the rung for this reason, and its tables for this vendor were rebuilt on 30 October against a ladder that has moved again since.

## The Refund Clause That Turned Into a Sliding Scale

FlyingPress never moved a price: $59 Starter, $109 Pro, $229 Business and $279 Unlimited on all seven readings, confirmed again on 2 November 2026.

Its exit condition moved twice. On 3, 10 and 17 October the page said, verbatim: *"We offer a 14-day free trial. No refunds needed - pay only if you continue after the trial."* On 24 October that sentence was gone, replaced by a scale. On 31 October the scale gained its last rung, and today it reads:

> "Refund requests: 100% within 24 hours of the charge; 80% after 24 hours and up to 7 days; none after 7 days."

A buyer who read the page in mid-October and bought in November was working from a policy that no longer exists. The price they compared is identical; what they can undo is not. Our [FlyingPress licence breakdown](https://xspeedcache.com/blog/flyingpress-pricing/) covers what the licence includes. The point here is that the clause moved while the figure sat still.

## One Page, Three Refund Windows

WP-Optimize Premium held $49.00, $99.00 and $194.00 across every reading. Read on 2 November 2026, its pages carry three statements about getting money back, and they are not one promise:

| Where on the site | What it says, verbatim | What it commits to |
|:---|:---|:---|
| Guarantee heading | *"TeamUpdraft plugins come with a 14-day money-back guarantee."* | An unconditional 14-day window |
| The sentence under it | *"If something goes wrong and we can't fix it, we'll offer a refund if it's requested within 14 days of purchase."* | A remedy, conditional on a fault going unfixed |
| Purchase FAQ | *"We consider all requests for refunds, particularly if made within the first 10 days after purchase."* | Discretion, on a softer 10-day window |

None of those three changed during the series: the first two were on the page when this engine recorded the conditional wording on 10 October 2026, and the FAQ answer on 17 October. A comparison table with a single "14-day refund" cell for this vendor is not wrong, exactly. It is unreadable as a commitment.

## The Entry Price Depends on Which Page You Land On

Object Cache Pro has not moved a figure. Its pricing page reads $95 a month or $950 yearly, Enterprise from $1,950 a month, on every reading since 17 October 2026. Its homepage says *"Starting at $79/month"*, and on 2 November that phrase appears three times and links to the page saying $95. This is the third consecutive reading in which the two disagree, first recorded on 24 October.

Which figure is right depends on a convention nobody publishes. This series treats the pricing page as canonical, which is why [our Redis object cache comparison](https://xspeedcache.com/blog/redis-object-cache-vs-object-cache-pro/) cites $95 and should not be "corrected" to $79. A buyer arriving from search quotes whichever page they landed on. Two further terms, read on 2 November: the subscription excludes the Redis instance, and a lapse gives 14 days' grace, after which, per the vendor's own FAQ, *"the object cache will stop working"*.

A headline price that excludes a separately billed component is a pattern with [its own name in the pricing literature](https://en.wikipedia.org/wiki/Drip_pricing), and the fair thing to say here is that each of these vendors states the exclusion on the same page as the price. FlyingPress names its CDN, Object Cache Pro names Redis, and ours names a Cloudflare add-on.

## Two Readings in This Series Were Wrong, and Both Were Ours

A series that never reports an error in its instrument is not careful.

**NitroPack's money-back window.** This engine recorded 30 days from 8 September 2026 and corrected it to 14 on 3 October, after a re-read found no "30" on the page. It was wrong in internal notes for 25 days and was caught before it reached a reader.

**WP Rocket's hosting exclusion, found while writing this article.** On 17 October the pricing FAQ was recorded as carrying a clause barring hosting companies from using the Multi licence for their clients. The readings on 24 and 31 October did not re-find it, and the 31 October note told future runs to treat the clause as gone and not to cite it. Read on 2 November 2026, it is on the page:

> "Yes, you can use our Multi License to set up WP Rocket on your customers' websites. [...] If you are a hosting company, you cannot use the Multi License for your clients. Instead, we offer a Reseller Program specifically for you."

The wording differs slightly from the 17 October record, which read "If you are offering hosting services". From outside a vendor's content management system there is no telling whether the sentence was edited, moved, or missed twice. What is certain is that "treat it as gone" was wrong, and an agency pricing fifty client sites on it would have been misled by us rather than by the vendor.

## The Five Fields That Decide What You Pay

Before buying any WordPress performance licence, read these five fields off the vendor's page and write the date beside them. Ten minutes for one vendor, and it survives every change this series recorded.

| Field | The question | Shapes it takes, read on 2 November 2026 |
|:---|:---|:---|
| **List price and unit** | What is the sticker, and what does a licence cover? | A site count, a one-time fee, or unlimited sites *"owned and operated"* by one entity |
| **Effective price** | What will the card be charged, and until when? | A countdown expiring 3 November 2026, a public code worth 10%, a three-year toggle worth 40% |
| **Entry condition** | Is there a trial, and does it charge automatically? | *"we'll charge your payment method automatically when your 14 days are up"*, or *"For licensing reasons, we don't offer a trial."* |
| **Exit condition** | What can be undone, and what does a lapse do? | A flat 30-day guarantee, a 100/80/none scale, a conditional remedy, or a 14-day grace period |
| **Renewal rule** | Does this price repeat? | *"will renew annually at the regular renewal price"*, against Perfmatters' *"15% discount automatically applied to all renewals"* |

![Five cards naming the fields to read before buying a licence: list price and unit, effective price, entry condition, exit condition and renewal rule.](https://xspeedcache.com/images/blog/vendor-term-volatility-fields.webp)

Those five fields reordered the ranking in our own three-year model. They are also the first thing a comparison table drops, because they do not fit in a cell.

## Our Own Row, Under the Same Columns

xSpeed Cache is ours and WPDeveloper is a Startise company, so our page goes into the matrix on the same terms as everyone else's. Read on 2 November 2026, [our pricing page](https://xspeedcache.com/pricing/) shows yearly licences at $29, $49, $99 and $199 against struck-through prices of $58, $98, $198 and $398, and lifetime licences at $79, $149, $249 and $399. A 14-day money-back guarantee covers all paid plans, and a Cloudflare Enterprise add-on is $5 a domain a month, billed separately, requiring a paid plan.

**We publish the one term nobody else in the series publishes.** Every other paid product here is a subscription, or a one-time fee with no published year-two answer. Ours publishes a lifetime price at every tier, *"One-time payment · no yearly renewal"*, which answers the renewal-rule field outright. NitroPack takes the opposite position: *"NitroPack does not offer a lifetime license because site optimization isn't a one-time task."* Both are defensible, and a buyer deserves both.

**Our page also still contradicts itself, and has for at least sixteen days.** The banner reads *"Get 50% OFF during Launch Week"*, the offer line reads *"50% off all paid plans for the first 6 months"*, and the yearly cards are labelled *"launch · 50% off for 6 months"* on plans the same page describes as *"Billed annually"*. A six-month discount on a twelve-month invoice is not a term a buyer can act on. Our own [three-year cost comparison](https://xspeedcache.com/blog/wordpress-caching-plugin-cost-comparison/) published that criticism of us on 6 October 2026, and it is unresolved today. Our row carries a standing presentation defect and no term moves. Being the vendor is no reason to grade our own page gently, and the fix is ours.

## Running This Across a Client Fleet Without Re-Reading Nine Pages Weekly

Nobody managing twenty-five client sites will run this protocol weekly, and nobody should. Run the five fields on the vendor you are about to buy, on the day the invoice goes out. The series above exists only because this engine publishes comparisons and must keep them true.

The second move is to take the licence off the renewal treadmill where you can. xSpeed Cache is our own caching plugin, and its lifetime tier exists so a fleet's performance licence stops being an annual re-decision: [Free vs Pro](https://xspeedcache.com/free-vs-pro/) sets out what the paid tier adds and the [80-capability comparison](https://xspeedcache.com/comparison/) places it against eight competitors. For agencies, [xSpeed Hub](https://xspeedcache.com/xspeed-hub/) manages a fleet from one place and the [agency use-case page](https://xspeedcache.com/use-cases/agencies/) covers that shape. If migration is the obstacle, the [migration documentation](https://xspeedcache.com/docs/migration/) imports settings from a competing plugin. Every comparison behind these readings sits in our [caching plugin hub](https://xspeedcache.com/blog/best-wordpress-caching-plugin/).

Run the free scan on your own site: [xspeedcache.com/scan/](https://xspeedcache.com/scan/)

## Common Mistakes Buyers Make When Reading a Dated Price

- 📌 **Trusting an undated price.** A price with no date is a claim about the past, and five figures here expire tomorrow.
- 📌 **Comparing list prices across billing models.** A one-time $49 and a $5-a-month-billed-yearly are not rungs on one ladder.
- 📌 **Assuming the renewal price is the purchase price.** One vendor's FAQ says it is not, one publishes a 15% renewal discount, and the rest say nothing.

## Frequently Asked Questions

### I quoted a vendor's price in a client proposal last week and it has already changed. Was I careless?

Not on this evidence. List prices were the most stable field in the series. The failure mode is quoting an effective price, which is what a promotion produces, with no expiry attached.

### My comparison spreadsheet says WP Rocket is $59.95 a year and the page says $5 a month. Which is right?

Both were right on their own dates. The page read $59.95 a year until 17 October 2026 and $5 a month billed yearly from 24 October, annualising to $60: a presentation move with a five-cent rounding, not a rise.

### I bought on the strength of "no refunds needed, pay only if you continue" and the page now shows a sliding refund scale. What happened?

FlyingPress replaced that sentence between the 17 and 24 October 2026 readings. The policy today is 100% within 24 hours, 80% up to seven days, none after seven. What governs a purchase is the policy in force at the charge, so keep a copy of the page you bought from.

### Why does Object Cache Pro say $79 on one page and $95 on another?

Its homepage has advertised *"Starting at $79/month"* while its pricing page has read $95 a month since at least 24 October 2026. This series treats the pricing page as canonical, and there is no telling which figure the vendor intends.

### I cannot fetch objectcache.pro with curl and get a 403. Is the vendor gone?

No. That host sits behind a firewall that refuses scripted requests, on every reading since 17 October 2026. A 403 there is an access control, not a dead vendor.

### My price scraper reported $520 and $2.97 on the same run. Which is the bug?

The whitespace. That page renders figures as `$ 5 / month` and `$ 2 .97`, so a match that does not normalise spaces inside numbers concatenates and splits them.

### How many readings does a price series need before it is worth anything?

More than one, and enough that a change can be dated. Seven readings over six weeks date each change to an eight-day window: enough to say a term moved, not which day.

### Is xSpeed Cache the cheapest of these?

Not on every axis, and the honest answer turns on the five fields rather than the sticker. Its $29 yearly launch price undercuts most of the field and its lifetime tier removes the renewal question, which no other product here publishes. Against that, WP Fastest Cache's one-time $49 needs no subscription at all, and Perfmatters' 15% renewal discount beats any annual licence without one.

### Which of these pages should I watch if I only watch one?

WP Rocket's. Six of the ten changes came from it, including both effective-price moves. The five pages that never moved are safe to read once.

### Can I reproduce this series myself?

Yes, and that is the intent. The nine pages are linked above, the five fields listed, and the tooling is `curl` with a browser user agent and a dated note per reading.

## Conclusion: Treat Every Price as a Reading With a Date

Ten changes across nine pages in six weeks, and not one a list price. The figures held while the trials, the refunds, the renewal rules and the promotions moved underneath them, which is the opposite of how comparison tables on this subject are built.

| If you are… | Read this field first | Because |
|:---|:---|:---|
| Buying one licence today | Effective price and its expiry | One headline price in this set ends 3 November 2026 |
| Budgeting for three years | Renewal rule | Only one of nine publishes a renewal discount |
| Buying for clients | List price and unit | One vendor bars hosting companies from its multi-site licence |
| Writing about any of this | The date you read it | Two readings here were wrong, and both were ours |

**What to do this week:**

1. Pick the vendor you are closest to buying and read the five fields off their page today, with the date beside each one.
2. Save the page you buy from, because the clause in force at the charge is the one that governs you.
3. Check whether any price in your own documentation carries a promotion.
4. Measure before you spend, and see which checks your site loses points on.
5. If the renewal treadmill is what you want gone, our lifetime tier is $79 at launch pricing with a 14-day money-back guarantee, and [Free vs Pro](https://xspeedcache.com/free-vs-pro/) shows what the paid engine adds.

Every figure above was read from the vendor's own page on the date stated. Re-read them before you spend anything, and if you keep a series of your own, publish its mistakes.