The Brief

The client — a Series A B2B SaaS company, anonymised at their request — had a well-funded growth team and was spending approximately $35,000 per month on paid search and LinkedIn advertising. Their site was receiving around 18,000 unique visitors per month. Their conversion event was a free trial signup.

At a 1.2% conversion rate, they were generating approximately 216 trial signups per month. Their trial-to-paid conversion rate was 28%, producing roughly 60 new customers per month. Their average contract value was $3,200 annually.

The brief was simple: improve conversion rate without touching the product, the pricing, or the acquisition channels. Everything had to come from the website itself.

Phase 1: The Diagnosis (Weeks 1–2)

The biggest mistake in CRO is skipping the diagnosis and jumping straight to implementing changes based on gut instinct or generic best practices. We have seen businesses add social proof to a site that already has strong social proof, shorten a form that is already optimal length, or rewrite hero copy that was already performing well — and waste weeks achieving nothing because they were fixing problems that did not exist.

Our diagnosis phase involves four inputs: quantitative data (where are people dropping off?), qualitative data (why are they dropping off?), heuristic evaluation (what does experienced analysis reveal?), and competitive benchmarking (what are comparable sites doing differently?).

Quantitative analysis

We set up Hotjar and reviewed two months of existing Google Analytics data. Three findings stood out immediately.

First, the homepage had an 82% bounce rate from paid traffic. Visitors arriving from Google Ads were leaving within 30 seconds at an extremely high rate. Second, the pricing page had a strong engagement rate — average time on page of 3 minutes 40 seconds — but only 8% of visitors who reached pricing were clicking the trial CTA. Third, mobile accounted for 41% of traffic but only 18% of conversions — mobile visitors were converting at less than half the desktop rate despite representing nearly half the audience.

Qualitative analysis

We ran an on-site survey using a single question displayed to 15% of visitors on exit: "What stopped you from starting a free trial today?" Over three weeks we collected 340 responses. The top four answers were: "Not sure if it works for my industry" (31%), "Couldn't find pricing information quickly" (24%), "Not sure what I get with the free trial" (19%), "Site was slow on my phone" (14%).

This was more valuable than three weeks of A/B testing. Four clear hypotheses, sourced directly from the people we were trying to convert.

Heuristic evaluation

Heuristic evaluation is experienced analysis against a framework — identifying friction points and conversion barriers through expert review rather than data. Our review flagged eight specific issues: the hero headline described what the product was rather than what it did for the buyer; there were no customer logos above the fold; the trial CTA said "Start Free Trial" with no secondary information about what "free trial" meant; the navigation contained 11 items creating decision paralysis; the mobile hamburger menu was 28px — below the 44px minimum touch target; pricing was hidden three clicks deep; there were no testimonials on any page except a dedicated testimonials page nobody visited; and the trial signup form asked for credit card details despite the trial being genuinely free.

Phase 2: The Hypothesis List

From diagnosis, we built a hypothesis backlog — 22 potential changes, each written in the format: "If we [change X], then [metric Y] will improve by [estimated amount], because [reason based on evidence]." We then scored each hypothesis on two dimensions: estimated impact (1–5) and implementation effort (1–5). Hypotheses with high impact and low effort go first. This is the ICE framework, and it keeps CRO programmes from getting bogged down in complex tests before the obvious wins have been taken.

The top five hypotheses from scoring were: rewrite the hero headline to be outcome-focused (high impact, low effort); remove credit card requirement from trial signup (high impact, low effort); add industry-specific social proof above the fold (high impact, medium effort); simplify navigation from 11 items to 6 (medium impact, low effort); add a FAQ section to the pricing page addressing trial scope and cancellation (medium impact, low effort).

Phase 3: The Tests and Results

Test 1: Hero headline — outcome vs feature framing

Hypothesis: Changing the headline from "The All-in-One [Product Category] Platform" to "Cut your [specific workflow] time by 40% — or your money back" will improve homepage-to-trial rate by 15%+, because outcome-framed headlines speak directly to the buyer's problem rather than the product's features.

Test setup: A/B test, 50/50 split, run for 21 days with statistical significance target of 95%. Control: original feature-focused headline. Variant: outcome-focused headline with a supporting sub-headline listing the top three customer-reported outcomes.

Result: Variant won. Homepage-to-trial click rate improved 34%. This was our biggest single test win of the programme. The lesson: most B2B software companies write their homepage headline for themselves — describing what the product is — rather than for their buyer — describing what the buyer gains.

Test 2: Remove credit card from trial signup

Hypothesis: Removing the credit card requirement from the trial signup form will improve form completion rate by 25%+, because credit card requirements are a significant friction barrier that signals potential charges, even when the trial is genuinely free.

Test setup: This required coordination with the product team to update the signup flow. Not an A/B test — a full implementation change tracked by comparing conversion rates before and after with statistical controls.

Result: Trial signup completion rate improved 41% in the two weeks following the change. This was the single highest-impact change in the entire programme, and it took 4 hours of product development time. The business had kept the credit card requirement because it reduced low-quality trial signups. The data showed the reduction in friction more than offset the reduction in lead quality — total qualified trials increased 29%.

Test 3: Industry-specific social proof above the fold

Hypothesis: Adding a strip of logos from customers in the visitor's likely industry (detected by IP-based company data or UTM parameters from targeted campaigns) directly below the hero will reduce homepage bounce rate by 10%+, because "not sure if it works for my industry" was the top exit survey response.

Test setup: A/B test. Control: existing homepage with customer logos in footer only. Variant: logo strip added below the hero headline featuring six customers, with a copy line "Trusted by [industry] teams at:" followed by logos.

Result: Homepage bounce rate from paid traffic reduced from 82% to 71% — an 11-point reduction. Trial conversion rate from the homepage improved 18%. The data confirmed the exit survey finding: visitors needed industry-relevant credibility early in the page, not just when they had already decided to scroll down.

Test 4: Navigation simplification

Hypothesis: Reducing navigation from 11 items to 6 will improve scroll depth and trial CTA click rate by reducing decision paralysis at the top of the page.

Test setup: A/B test. Six items retained: Product, Pricing, Customers, Blog, Login, Start Free Trial (the CTA). Five items removed or consolidated into dropdowns: Resources, Company, Partners, Integrations, Changelog.

Result: Inconclusive. The 6-item navigation showed a marginal improvement in CTA click rate (+4%) but the result did not reach statistical significance in the 21-day window. We kept the simplified navigation because it improved the overall page aesthetic and the directional trend was positive, but we cannot attribute conversion improvement to it with confidence.

Test 5: Pricing page FAQ

Hypothesis: Adding a FAQ section to the pricing page addressing "what is included in the free trial", "do I need a credit card", "can I cancel anytime", and "what happens when the trial ends" will improve pricing-page-to-trial rate from 8% to 12%+.

Test setup: A/B test. Control: existing pricing page. Variant: FAQ section added below pricing tiers, above the footer CTA.

Result: Pricing-page-to-trial rate improved from 8% to 14.3%. Statistically significant. The FAQ addressed the exact objections raised in the exit survey. The lesson: answer the questions your visitors are already asking, at the moment they are most likely to be asking them.

Tests 6–11: Secondary tests with mixed results

We ran six additional tests during the programme. Two won clearly: adding a "no credit card required" line directly below the trial CTA button improved trial page completion rate 12%; adding a sticky trial CTA button on mobile improved mobile conversion rate 22%. Two lost: adding urgency messaging ("Join 430 companies who signed up this month") reduced conversion slightly — likely because it felt manipulative to a sophisticated B2B audience. Adding a chatbot to the pricing page produced no statistically significant change and was removed. Two were inconclusive and deprioritised.

The Results: 90 Days Later

The compound effect of the winning changes was larger than the sum of individual test improvements — which is typical in CRO, because visitors do not experience one change in isolation, they experience the full site. After 90 days and seven implemented changes:

  • Site-wide conversion rate: 1.2% → 2.6% (117% improvement)
  • Monthly trial signups: 216 → 468 (same traffic budget)
  • Monthly new customers at 28% trial-to-paid: 60 → 131
  • Monthly revenue from new customers at $3,200 ACV: $192,000 → $419,200
  • Incremental annual revenue attributable to CRO: approximately $2.7 million
  • Cost of the CRO programme: $14,400 (PixelForge engagement fee)
  • ROI: approximately 185x in the first year

What Applies to Your Website

The specific changes that worked for this client may not be the right changes for your site — CRO is inherently site-specific. But the process applies universally, and three principles from this programme apply to almost every business website we encounter.

First: your hero headline is almost certainly about your product rather than your customer's outcome. Test a version that leads with the specific result your best customers experience. The improvement is reliably among the largest single changes available on most sites.

Second: there is at least one form on your site with a field that is creating friction without adding value. Audit every form. For each field, ask: "Is this information required to complete this action, or are we collecting it for our convenience?" Remove every field that answers "our convenience."

Third: your visitors are leaving with unanswered questions. An exit survey — a single question, non-intrusive, shown to a small percentage of exiting visitors — will tell you what those questions are within two to three weeks. The answers are almost always surprising, consistently actionable, and worth more than any amount of educated guessing about what to test.

If you want to run this process on your site, book a free 30-minute call. We will look at your current setup and tell you where we would start — based on what we actually see in your data, not a generic checklist.