New format
Conversion rate optimization: how to measure it and what to test
Conversion is the share of visitors who complete a target action: a lead, a purchase, a call, a signup. It’s measured in percent and improved with focused edits — not “more ad budget.”
Below: the formula, the funnel, UX and content hypotheses, and A/B tests. “Average niche conversion” figures from old roundups are rough cues — not your KPI.
What conversion is
Conversion rate (CR) is target actions divided by visits over a period. You define the goal: order, lead, demo, subscribe.
Also watch micro-conversions: phone tap, finishing a key block, add to cart — they show where the path breaks.
Example:
- 3,000 landing visits
- 150 signups
- CR = 150 ÷ 3,000 × 100% = 5%
Strategy before button tests
Without knowing the audience and offer, a CTA color change rarely saves you. First: who arrives, what the ad or search promise is, whether the landing matches.
In analytics (GA4, Yandex Metrica, and similar) watch sources, devices, exit pages, goals. Sticky content should lead to a commercial step without bait-and-switch.
If pages don’t hold:
- refresh meaning and proof
- fix broken links and forms
- clarify the CTA
- cut visual noise
Funnel and product
Classic path: awareness → interest → comparison → action. On a site that’s the road from the first screen to a lead or payment. A weak link kills everything below.
A 5% discount hasn’t been a universal lever for a long time. Stronger: clear value, a demo or estimate, guarantees, social proof, a fast reply.
Align product and audience:
- offer = audience pain
- price and timeline clear before the form
- a next step without “fill 12 fields”
Hypotheses: forms, CTAs, speed, trust
Heatmaps and session recordings show where people click and drop. Next — a hypothesis and a measurement.
Form practice:
- name + contact + gist
- hints and validation errors
- thank-you with a real SLA
Typical CR levers
| Zone | What to try | Risk |
|---|---|---|
| Form | Fewer fields, clear reply window | Weaker lead qualification |
| CTA | Benefit copy, contrast, repeat by the offer | Noise if too many buttons |
| Speed | Compression, cache, mobile layout | Needs a developer |
| Trust | Reviews, cases, guarantee, contacts | Fake reviews worse than none |
| Content | Lists, facts, fewer junk banners | Cutting value just to be “shorter” |
A/B tests — and when to skip them
Split: two variants of one change, split traffic, measure the goal. Change one thing at a time. A week “by eye” with no volume isn’t a test.
With low traffic, first remove obvious barriers (broken form, five-second LCP, no phone). Then scale experiments.
Takeaways
Conversion optimization is a loop: goal → data → hypothesis → edit → measure. Count CR honestly, fix the funnel and trust, don’t copy someone else’s percentages. Ads without a converting landing only make the click more expensive.
FAQ
How do you calculate conversion?
Target actions ÷ visits (or sessions) × 100%. Fix what counts as the goal and measure in one tool (analytics / GA4 / CRM).
What conversion rate is “normal”?
It depends on niche, traffic, and deal complexity. Compare yourself to yourself: week to week, channel to channel — not to someone else’s “9% on social.”
Should you fix traffic first or conversion?
With expensive traffic and a leaky funnel, fix conversion first. With tiny traffic A/B is pointless — grow volume and remove obvious barriers in parallel.
Is an A/B test always needed?
For small edits with low traffic, before/after and common sense often suffice. Run A/B when you have volume and one clear hypothesis.
Do forms with many fields cut conversion?
Yes — extra required fields. Capture the minimum for first contact; details after a manager replies.
Traffic is fine but leads are thin?
We’ll fix the funnel: form, CTA, speed, and trust — measured as CR, not vibes.
Discuss the task
