B2B SaaS pricing experiment: Eight weeks, four hypotheses, real case

If you suspect your pricing is throttling growth but aren’t sure which lever to pull, this case will tell you where to start. In the first quarter of 2026 we ran an eight-week pricing experiment program with one of our B2B SaaS clients. The client stays anonymous in this post — the operating team preferred not to put the name in writing, and we respect that — but the sector, scale, and lessons are useful enough on their own. The product is a B2B platform for operational excellence, with 24+ months of proven product-market fit and quarterly MRR growth in the 8-12% range. The question was simple: is the pricing structure slowing the growth or could it speed it up?

Below is the program from start to finish: starting state, four hypotheses, the order of the four experiments, results, and net impact. The method is portable; the numbers are specific to this client but useful as a reference for B2B SaaS at similar scale. We ran the program as part of strategy and insights work; pricing programs of this shape typically run 6-10 weeks.

Starting state — what we were changing

Eight weeks before launch, the picture was three plan tiers and a flat per-seat price:

• Starter: $69/seat/month, capped at 10 seats, reduced feature set.
• Professional: $89/seat/month, unlimited seats, full product.
• Enterprise: “Talk to us” — no real Enterprise contract existed, most leads got routed to Professional.

Trial-to-paid conversion sat at 4.2% (below the sector benchmark; target was 6%+). Annual upgrade rate was 18% (also low). NPS sat at 42 — solid but not great. Roughly 75% of ARR came from Professional, Enterprise was effectively unsold, and Starter was the only option for a tight mid-small segment.

The pre-program diagnostic surfaced two findings. One: the Professional seat price was reasonable for most clients, but the “10 to 50 seats” band kept producing ad-hoc discount requests at sales close. Two: there was no free tier, and the trial was capped at 14 days — too short for a buyer running a careful evaluation. Those two findings drove half of the hypothesis list.

Four hypotheses

At the end of the diagnostic, four hypotheses were written down and prioritized. Each had a success criterion and a rollback plan defined upfront — pricing experiments include changes that, if mishandled, damage the customer base.

• H1: Per-feature unbundling. If we split a few Professional features (advanced reporting, API access, custom workflow editor) into separate paid modules, does ARPU rise? Expectation: rises, because larger customers buy more modules and smaller ones enter at a lower core price.
• H2: More aggressive annual discount. The existing annual discount was 15% (two months free). Move it to 25% (three months free); how much does annual upgrade rate change? Expectation: 30%+ relative lift on annual conversion, with cash collection and net retention up even if MRR holds.
• H3: Free tier. Add a Free tier capped at 5 seats and 200 actions per month. Does this raise trial sign-up and net paid conversion? Expectation: sign-ups rise sharply, paid conversion drops in percentage terms but holds or grows in absolute count; net positive needs a 5-10% conversion floor.
• H4: Real Enterprise tier. Replace “Talk to us” with an actual Enterprise package — minimum 50 seats, custom SLA, priority support, SSO, audit log. Does this win net new logos? Expectation: 1-3 deals close, sales-led motion required.

For each hypothesis, the duration, segment, success threshold, and rollback protocol were written into the program contract. Product, marketing, and sales teams signed off before launch.

Experiment 1: Per-feature unbundling — failed (-18% conversion)

The first experiment was the boldest. Over a three-week window, half of incoming trials saw the existing flat Professional plan; the other half saw a “Modular Professional” structure: $59/seat/month core + $19 advanced reporting + $29 API + $19 custom workflows. Total with all modules: $126 — nominally more expensive, but framed as “buy only what you need.”

The result landed hard. Trial-to-paid conversion was 4.1% in the control and 3.4% in the modular cohort. An 18% drop. ARPU did not rise either: most modular buyers picked “core plus one module,” coming in around $78/seat/month — below the flat $89. Three weeks in, we killed the hypothesis and pulled the modular structure.

The lesson: B2B buyers do not pay for pricing complexity. During the sales call, “should we add this module, what does it unlock” creates friction; sales cycles lengthen; decisions get pushed. Modular pricing can work in B2C (Spotify-Hulu style) but B2B buyers want “all-in, predictable.”

A note on rollback: 11 customers who had bought during the modular window were upgraded to flat Professional at no extra cost. For them, free value increase; for us, avoiding negative reputation damage. Pricing experiments should always have a “loss compensation” plan written upfront.

Experiment 2: Annual discount 15% → 25% — success (+34% annual upgrade)

The second experiment was more cautious: at renewal time, existing customers were offered 25% off for switching to annual (up from 15%). New trials saw the same offer. Measurement window: four weeks of annual upgrade rate plus net 90-day ARR.

The result was clean. Annual upgrade rate moved from 18% to 24.2% (+34% relative). Net 90-day ARR was up 22%. Cash collection — annual upfront — added roughly $340K extra in the window (meaningful working capital for a mid-size SaaS). Net dollar retention rose from 105% to 109% per quarter: annual customers churned 3% lower than monthly.

Main risk: does an aggressive annual discount frame the monthly plan as expensive? Data said no — monthly plan growth held. But the precondition is that the product is stable enough for buyers to commit annually. In an early-stage SaaS, an aggressive annual discount creates upfront payment plus refund risk; this client had 24+ months of stability so the lever worked.

Operational lesson: the discount alone was not the whole story. When we paired it with an email campaign offering “two free quarters of advanced analytics for annual upgraders,” upgrade rate added another 4 percentage points. Pricing experiments rarely succeed as pure pricing changes; messaging, channel, and bonus packages move together.

Experiment 3: Free tier — success (+67% trial sign-up, +9% net paid conversion)

The third experiment produced the largest impact and the most internal debate. The new Free tier capped at 5 seats, 200 actions per month, core feature set, branding fixed. Customers hitting limits saw a “Upgrade now” CTA in the product; after 30 days of inactivity, the team got a personal email.

Two weeks after launch, weekly sign-ups moved from 312 to 521 (+67%). The first concern was whether Free would cannibalize Starter. It did not — Starter sign-ups held flat (about 85/week) and Free-to-Starter upgrades came in at 38/week. Net paid conversion moved from 4.2% to 4.6%; marginally higher in percentage but in absolute terms about 45 new paid customers per week, equal to roughly $3,150 new MRR/week or $163K new ARR/year.

The unexpected upside of the Free tier landed elsewhere: in-product social signal. Roughly half of Free users worked at companies that already had paid users on the platform. “Company X has 4 free users plus 2 paid users, how do we consolidate” became a high-quality outreach prompt for the sales team. The intersection of PLG and sales-led motion — six weeks after Free tier launched, the outbound team rebuilt its target account list around companies with Free users.

Cost side: Free tier consumes backend storage, compute, and support overhead. The math: 1,000 Free users = roughly $1,200 fixed monthly cost. Net paid conversion above 5% keeps it profitable; below 5%, revisit. Sitting at 5.2% currently, on the watchlist.

Experiment 4: Enterprise custom — mixed (3 deals, sales-led motion needed)

The fourth experiment landed last in the eight weeks but represented the largest contracts. The generic “Talk to us” was replaced with an actual Enterprise package: 50-seat minimum, $79/seat/month (lower than Professional, made up in volume), SSO, audit log, priority support, 99.9% SLA, dedicated success manager. Public pricing page now reads “starts at $4,000/month.”

Eight-week numbers: 19 Enterprise leads, 12 demos, 5 proposals, 3 deals closed (combined $14,700/month or roughly $176K ARR). Four more deals still in pipeline. Fastest deal was 5 weeks; slowest 7.5. Close rate landed at 16% (3/19) — average for the sector but not strong.

We marked the experiment as “mixed” because two things surfaced:

• Enterprise sales requires sales-led motion. The current customer success team is one person; sales is three SDRs and one AE. Enterprise needs its own channel — outbound, networking, conferences. PLG flow does not carry it.
• Discovery itself can be a paid step. Enterprise buyers asked for 3-4 demos plus 2 technical Q&As plus 1 PoC. The PoC took roughly 10 hours/week of client team time. Free pre-sales effort lengthens payback; an alternative is a “paid pilot” — for example, $5K for a 4-week pilot that converts to annual on success.

Decision: keep the Enterprise tier, but in Q2 2026 ship a separate Enterprise sales playbook — outbound, paid pilots, a dedicated AE. That is a separate engagement, outside the 8-week program scope.

Net impact and program-level lessons

Net impact across the eight weeks:

• MRR growth: +23% (composition: H2 +6%, H3 +18%, H4 +4%, H1 -5%, net +23%).
• ARR growth: +41% (annual upgrade lift plus Enterprise commitments).
• Trial sign-up: +67% (driven entirely by H3).
• Trial-to-paid conversion: 4.2% → 4.6% (+9% relative).
• NPS: 42 → 47 (Free tier surprisingly lifted NPS — engaged non-paying users score high).

Three program-level lessons that look sector-independent:

• B2B buyers pay for clarity, not complexity. H1 (modular unbundling), however well-modeled mathematically, slowed the buying decision and turned negative. A flat all-inclusive plan is “I can sign without thinking” for the buyer.
• Annual discount is underrated. Compound effect on cash flow, retention, and net dollar retention; most B2B SaaS does not pull this lever hard enough.
• Free tier, designed correctly, becomes a Trojan horse. It generates in-product social signal that enriches outbound sales lists. Do not evaluate it on conversion metric alone.

A note on the limits of this program

The 8-week experiment carries one final reminder: pricing experimentation is not “find the right price.” Pricing is a minor adjustment for a SaaS with proven product-market fit, and a dangerous lever for one without. The client we worked with had 24+ months of product stability; the same program for a SaaS still hunting PMF would not have produced these results. Product first, price second — flip the order and the price experiment masks a product gap.

We will publish a fuller anonymized write-up on our case studies page once the client signs off (estimated Q3 2026). If you want to run a similar pricing program, or a wider strategy review of your SaaS, reach out — bring your starting state and current metrics to the discovery call, and we agree the priority list of hypotheses together.