Conversion rate optimization for a small site: what to fix first

Conversion rate optimization is the practice of increasing the share of visitors who do the thing you want — enquire, sign up, buy, book. It is usually presented as a testing discipline, and for sites with heavy traffic it is. For a small business site with a few thousand visitors a month, that framing is actively unhelpful: you will not reach statistical significance on a button colour this decade.

What works at small scale is different and less glamorous. Fix the things that are obviously broken, make the offer clearer, remove friction from the form, and only then think about testing. Almost every small site has weeks of work available before testing becomes the right tool.

The order to work in

• Measurement first. If you cannot see how many people arrive, how many convert, and where they leave, everything after this is guesswork.
• The offer. The most common conversion problem is not the page — it is that the thing being offered is not what the visitor wants, or the price and terms are unclear.
• The first screen: what this is, who it is for, and what happens next, answerable in five seconds without scrolling.
• The call to action — one primary action per page, stated as what the visitor gets rather than as a verb like "submit".
• The form. Every field removed raises completion; every field added must earn its place.
• Speed and mobile. A page that takes six seconds on a phone converts badly regardless of copy.
• Trust: prices, real contact details, a named person, evidence you have done this before.
• Only then testing — and only on the pages with enough traffic to learn something.

Before optimising anything, watch five people use the page — colleagues who have not seen it, a customer, anyone. Watching someone hesitate for four seconds looking for the price teaches you more than a month of small tests, and it costs an afternoon. Nearly every small site has at least one problem that is obvious on the first watch and invisible in the analytics.

The changes that usually move the number

1. Say what you do in the first sentence, in the words customers use, and put it above the fold.
2. Publish prices, or explain honestly why you cannot. Hidden pricing is the most common reason B2B visitors leave without contacting anyone.
3. Cut the form to the fields you genuinely need to respond. Name, email and one context field will outperform eleven fields every time.
4. Replace "Submit" with what happens: "Get the quote", "Book the call".
5. Set expectations at the point of action — what happens after they click, and how quickly you reply.
6. Add one piece of specific evidence near the call to action: a named client, a number, a result. Generic testimonials do very little.
7. Remove competing links from a page with one job. Every alternative path is a leak.
8. Check the page on a mid-range phone on mobile data, not on your laptop.
9. Fix the thank-you page, which almost everyone forgets and which is the best place to set expectations.

When testing is worth it

A rough rule: you need enough conversions per variant per test — commonly cited as a few hundred — before a result means anything. A page with forty conversions a month cannot A/B test its way anywhere; the honest options are to make bigger changes and compare periods, or to accept qualitative evidence. Bigger changes are also where the larger gains are: rewriting an offer beats testing a headline, and for a small site the sequencing matters more than the sophistication.

In Ettex, the page itself is built in Ettex Sites — a builder with sections, a live canvas, templates and one-click publishing on your own domain, which matters here because the ability to change a page in an afternoon is what makes iteration possible at all. The forms on the page feed Ettex Forms, where submissions land timestamped in one searchable inbox with a live summary, so "how many enquiries did we actually get" is a number rather than an impression. Enquiries can be tracked through Ettex CRM, the follow-up goes out from Ettex Mail, and the record of what you changed and when belongs in Ettex Docs — without it, you cannot attribute a change in results to anything.

Plainly: Ettex has no CRO tooling. There is no A/B testing, no analytics, no heatmaps, no session recording, no funnel reporting and no personalisation. You will need analytics from elsewhere to measure any of this. What Ettex gives you is a page you can change quickly and a form whose submissions you can count — which, at small scale, is most of what the work actually requires.

Why CRO efforts stall

• Testing started before measurement exists, so results cannot be read.
• A/B tests run on traffic far too small to produce a conclusion, then acted on anyway.
• Button and colour tests while the offer itself is unclear.
• Forms that ask for phone numbers, company size and budget before any trust exists.
• No prices, so visitors self-select out silently.
• Multiple calls to action competing on one page.
• Changes made in batches, so nothing can be attributed.
• No record of what changed when, which makes the whole history unreadable six months later.

Conversion rate optimization at small scale is sequencing, not statistics. Measure first, fix the offer, shorten the form, make one action obvious — and watch five people use the page before you change anything.