How we lifted demo bookings 58% by adding live chat to our pricing page

We ran an A/B test on the RunPivot pricing page using our own platform. The variant — a single Tawk.to live chat widget — beat the control convincingly. Here's the full breakdown: what we tested, what happened, and why we think it worked.

The problem

Our pricing page was converting at 3.8% into demo bookings. Decent, but we had a hunch that late-stage visitors — the ones scrolling through plan comparisons — were hitting a silent wall. They had questions. "Does this work with my SPA?" "How is MTU counted?" "What happens at the end of the free trial?" The FAQ answered some of it, but nobody was bothering to scroll back up.

In CRO, the pattern is well-known: intent is highest right at the pricing table, and the cost of an unanswered question at that moment is a lost deal.

The hypothesis

If we placed a lightweight live chat widget on the pricing page only — not sitewide — visitors evaluating plans could ask one quick question and stay in flow. We expected this to shift more visitors from "I'll think about it" into "Book a demo" before the tab got closed.

We chose Tawk.to because it's free, loads fast, and we wanted to see whether the presence of chat itself — not a fancier tool — was the lever.

The test

50/50 traffic split. Ran for 14 days across roughly 4,200 visitors. The primary metric was demo bookings from the pricing page; we also tracked scroll depth and time on page as guardrail metrics.

The results

Why it worked

1. Intent-matched placement

Chat on a homepage is noise. Chat on a pricing page is a handrail. The visitors already have purchase intent — they're there to evaluate cost, fit, and risk. A small affordance that says "someone is here if you have a question" pattern-matches to the moment.

2. Questions get answered before the tab closes

Most pricing-page exits are silent. The visitor has an objection, doesn't voice it, and leaves. Live chat converts a subset of those silent exits into a 30-second conversation. Even unanswered chats (where the visitor doesn't engage) seem to reduce abandonment — the presence of the widget itself signals a staffed, real product.

3. Trust via availability

For early-stage SaaS, "is anyone home?" is a real question in the buyer's head. A visible support surface is a cheap way to answer it. This is the same psychological beat as social proof — we're just answering it with presence instead of testimonials.

What we're testing next

RunPivot's thesis is that SaaS websites should be tested continuously, not annually. So the pricing-page test isn't the end of the story. Variants queued up next:

1. Proactive chat trigger after 20 seconds on the pricing page vs passive widget.
2. "Talk to a founder" framing vs generic "Chat with us" as the CTA label.
3. Inline FAQ accordion inside the pricing table vs. chat — same problem, different UX.

We'll write each of these up the same way. The point isn't that live chat is a silver bullet — it's that there's almost always a meaningful lift hiding in a well-scoped test, and running them continuously compounds.

How to replicate this

If you want to run the same test on your own pricing page, here's the setup:

1. Sign up for a free Tawk.to account and grab the embed snippet.
2. In RunPivot, create a new test scoped to your pricing URL.
3. Add the Tawk.to snippet to the variant only (RunPivot injects it without touching your code).
4. Set demo booking (or trial start) as your primary goal, 50/50 split, run until you hit significance.

End-to-end, the setup takes under ten minutes.

Takeaways

1. Intent beats cleverness. The winning change wasn't novel — it was placed where buyer intent was already high.
2. Silent objections are your biggest leak. Anything that lowers the cost of asking a question is worth testing.
3. Small, boring tests compound. +58% from one widget swap. Stack four or five of these in a year and the curve bends.