Pricing page edits are the highest-leverage change on a self-serve site, and the easiest place to misread analytics. You ship new copy Friday, revenue dips Saturday, panic Sunday, rollback Monday, never knowing if the dip was price, layout, traffic mix, or that one person who always refunds.
RPV after a pricing change only works if you pin the before/after windows and hold traffic sources as steady as real life allows. Part of the revenue per visitor guides; read the Main guide for definitions first.
What changes when you touch pricing
Three different events get conflated:
- List price change (numbers on the page).
- Packaging change (features move between tiers).
- Presentation change (layout, annual toggle default, social proof).
Each hits RPV differently:
- Price up → RPV might rise per visitor even if conversions fall.
- Price down → RPV might fall per visitor even if volume rises.
- Annual default → cash timing shifts; RPV on cash basis can spike while MRR story lags.
Name which event you shipped in one sentence in your log.
Pin ranges like a lab notebook
Before deploy, create two pinned ranges in your analytics (or calendar blocks if you are on spreadsheets):
- A: 14–28 days pre-change (baseline).
- B: day 0–14 post-change (early reaction).
Do not peek at day 2 and declare victory. Early buyers are often warm audience who already decided.
Extend B to 28 days when traffic is under 10k, see RPV on thin traffic.
Segment /pricing landing RPV separately
Site-wide RPV blends home tourists with pricing-intent visitors. After a pricing change, prioritize [RPV by landing page](/blog/rpv-by-landing-page-for-one-core-offer) filtered to /pricing and to campaign landers that point to checkout.
If only pricing-lander RPV moves while home RPV is flat, you likely changed buyer perception, not top-of-funnel.
Conversion rate and RPV can disagree, on purpose
Example after a 20% price increase:
- Visitors: flat
- Conversion rate: down 15%
- Revenue per visitor: up 5%
That can be a win for cash if churn stays stable. Founders panic on conversion alone.
Write both numbers in your review:
CR: 2.1% → 1.8%RPV: $0.52 → $0.55
Decision belongs to your margin and support load, not Twitter analytics culture.
Traffic mix confounds price tests
If you launch on Product Hunt the same week you raise prices, site RPV tells a muddy story. Either:
- Delay price change a week after launch traffic fades, or
- Tag launch traffic with UTMs and compare non-launch segments only.
There is no perfect isolation on a small site. Honest pinning beats pretending you had a control group.
Annual toggle and RPV timing
Switching default from monthly to annual makes cash RPV jump on the week toggles flip, even if LTV is unchanged. Note the toggle date separately.
For subscription businesses, pair RPV with refund rate 30 days out on the new default plan.
Layout-only changes
Moving testimonials above the fold does not change price but can change who converts. Measure:
- RPV on
/pricing - Scroll depth if you have it
- Support tickets mentioning “could not find feature X”
Layout wins show up as higher RPV at the same price point.
Rollback criteria (decide before ship)
Pick rollback triggers in advance, for example:
/pricingRPV down >25% for two consecutive 7-day windows with stable traffic.- Refund rate doubles vs baseline.
- Checkout errors increase (technical regression).
Avoid rollback on one bad day or one angry tweet.
Worked scenario
Baseline 28d: 5,000 visitors, $2,250 revenue → $0.45 RPV.
Post-change days 1–14: 2,400 visitors, $1,200 revenue → $0.50 RPV.
Conversion down, RPV up, price increase likely working for cash. Wait days 15–28:
- If RPV holds → keep.
- If RPV reverts while complaints rise → reconsider packaging, not just number.
Communicate price changes to returning users
Returning visitors may hit /pricing expecting old numbers. Spike in direct + support mail is common. RPV may dip for a week while trust resets.
Optional: one changelog email with UTMs so you can see email-led pricing visits separately in the pin.
International pricing
If you add EUR pricing display, RPV by country prevents comparing US week to blended week. Segment country or currency in the pin.
What analytics will not tell you
- Whether you underpriced emotionally.
- Whether enterprise leads were scared off (they may never hit self-serve checkout).
- Long-term churn, RPV is not LTV.
Follow revenue retention in your billing tool; use RPV for front-door efficiency.
Checklist the week you ship
- Log change type (price / packaging / layout).
- Pin baseline and post windows.
- Watch
/pricingRPV, not only totals. - Hold major campaigns constant if possible.
- Review at day 14 and day 28, not day 1.
Pricing work is stressful enough without metrics that change definitions mid-crisis. Pin the dates, read RPV with conversion beside it, and let small samples finish their sentence.
Documenting the before state (screenshots are not enough)
Save the old pricing HTML or a PDF export, the previous Stripe price ids, and the baseline RPV pin in the same note. Screenshots of charts go stale when you change date pickers. A pinned range with a label like pre-price-2026-01 survives staff changes and fuzzy memory.
If you use feature flags to roll out pricing to 50% of visitors, analytics must record variant on the session or you are blending two price lists into one meaningless RPV. Either ship to everyone at once on a small site, or instrument the flag.
Competitor comparison tables on pricing
Adding “us vs them” tables changes perceived value without changing SKU price. RPV can rise from better framing, not from discounting. Log “presentation-only” in your ship note so you do not confuse a framing win with a price test when reviewing lander RPV.
Moving pricing URL (/pricing → /plans)
Path moves split history unless you merge paths in analytics. Pin pre-move and post-move separately; do not compare blended 28d. Update internal links and saved funnels the same day or path analysis breaks. Note the redirect map in the same pin comment so RPV reviews stay legible six months later.
Support tickets as a lagging RPV signal
RPV can look fine for seven days while tickets ask “where did the starter tier go?” Tag pricing-related tickets that week. If ticket volume doubles and /pricing RPV is flat, you may have confused existing users without gaining new payers, packaging problem, not price level.
Free trial changes vs list price changes
Moving from 14-day to 7-day trial does not change cash RPV until trials convert. Watch signup volume immediately and checkout_completed two weeks later. Mixing trial changes with list price changes in one deploy is how founders lose the plot entirely; stagger changes by at least one full 28-day RPV window when possible.
After you stabilize
Once post-change RPV matches or beats baseline for a full pin, update your public case study numbers and marketing site quotes if they mention old pricing. Stale $9 screenshots next to $12 checkout hurt conversion more than any A/B tool will fix, and they show up as weird /pricing RPV dips when visitors bounce at mismatch.