- Website revenue is visitors × conversion rate × average order value × repeat purchases. To increase sales, find the weakest of the four factors and work on that one first.
- Because the factors multiply, a 10% gain in any one of them lifts revenue by 10%. That is arithmetic, and it shows why a balanced view of all four beats working on traffic alone.
- Every change is a hypothesis. An A/B test tells you which change really raised sales, because both versions run at the same time under the same conditions.
- The page ends with a 30-day plan: measure in week 1, choose a lever in week 2, run the first test in weeks 3 and 4.
To increase sales on a website, work on four factors: the number of visitors, the share of them who buy, the value of each order and how often customers buy again. Find the weakest factor, change one thing there, and measure whether sales really went up.
Many guides start with more traffic or with a list of design tips. This page starts with the revenue formula, because it shows where your next hour of work has the biggest effect. For each of the four levers you get the first step, a link to the detailed guide, and the metric that tells you if it worked. The last two sections explain how to prove which change raised sales and give you a 30-day plan.
The revenue formula: visitors × conversion rate × order value × repeat purchases
Website revenue is the product of four factors: how many visitors arrive, which share of them buys (the conversion rate), how much each order is worth (the average order value), and how often the same customer buys again (repeat purchases). Every change you make to a website moves one of these four.
people who arrive
share who buy
spend per order
orders per customer
your sales
Because the factors multiply, a 10% gain in any one of them lifts revenue by 10%, and small gains in several factors compound. That is plain arithmetic, with no benchmark behind it. It also explains why the biggest lever is rarely the one you are working on today: the factor that is weakest for your business is the one with the most room.
Use this table to find your bottleneck and the page that covers the first step:
| Lever | Typical measure | What to track | Where to read more |
|---|---|---|---|
| 1. Visitors | Search visibility, ad targeting, email and social traffic | Sessions and new users by source, in your analytics | Website optimization |
| 2. Conversion rate | Clearer product pages, shorter checkout, more trust signals | Orders divided by sessions, per device | Why your shop is not converting |
| 3. Order value | Cross-selling, bundles, a free shipping threshold | Revenue divided by orders | Increase average order value |
| 4. Repeat purchases | Follow-up emails, customer accounts, reasons to return | Share of customers with a second order | Covered below |
A quick way to pick the bottleneck: if few people reach your site, start with lever 1. If many arrive and few buy, start with lever 2. If people buy but each order is small, start with lever 3. If customers buy once and never return, start with lever 4. Most shops find their first answer in the funnel report of their analytics tool.
Lever 1: more qualified visitors
More visitors only help if they are the right visitors. A campaign that sends people who never wanted the product raises the visitor count and lowers the conversion rate at the same time, so the revenue formula does not improve.
Keep this lever short and specific:
- Match the promise to the page. The ad or search snippet and the landing page should answer the same question.
- Check the quality of each source. Compare conversion rate and order value for each source, next to the number of sessions.
- Fix the basics of findability. Titles, descriptions, page speed and a clear structure are the foundation. The guide to website optimization covers SEO, technology, UX and accessibility.
This page stays brief on traffic on purpose. Acquisition has its own channel guides. What follows concentrates on the three levers you can influence on your own site.
Lever 2: convert more of your existing visitors
Conversion rate optimization (CRO) is the work of turning more of your existing visitors into buyers. It is the lever with the shortest path, because the visitors are already there. To put your own rate in context, compare it with the e-commerce conversion rate benchmarks.
Start with the questions a visitor must be able to answer:
- What is this, and is it for me? The headline and the main image answer it on the first screen.
- What does it cost in total? Shipping and fees are visible before the cart.
- Can I trust this shop? Reviews, return terms and contact details are easy to find.
- Is buying easy? The cart and checkout have no detours, and guest checkout is available.
If you are not sure where visitors leave, the diagnosis in Why your online shop is not converting walks through the buying path stage by stage. To get concrete test ideas for a page, enter its URL into the AI CRO Audit. It returns insights, recommendations and ten test ideas you can use directly as A/B tests. For the method behind it, see conversion optimization.
Lever 3: raise average order value
Average order value (AOV) is revenue divided by the number of orders. Raising it means each buyer spends more per order, without you needing more buyers.
The usual tactics are cross-selling, up-selling, bundles, a threshold for free shipping and recommendations in the cart. Each has a risk: an offer that feels pushy can lower the conversion rate. Always watch both numbers together.
The page How to increase average order value lists nine tactics with placement, risk and what to measure for each.
Lever 4: bring buyers back
A customer who has already bought once knows your delivery, your product and your service. Bringing that customer back costs less effort than winning a new visitor, and every extra order from the same person raises the fourth factor in the formula. These steps work without any special tool:
- A useful follow-up email. Send care tips, usage ideas or a note when a consumable is likely to run out.
- A newsletter with a reason to open it. Share new arrivals, restocks or advice. Ask for the address at a moment when the visitor sees the benefit.
- A customer account that saves effort. Order history, saved addresses and easy reordering give a returning buyer a faster path.
- Clear service after the sale. Easy returns and quick answers make a second purchase more likely. Ask buyers for a review a few days after delivery.
- A look at your own data. Find out how many customers order twice. If you have never measured it, that measurement is your first step on this lever.
Keep the promises small and true. Send a reminder at the right moment, and use discounts sparingly.
How to know which change actually raised sales
Every change in the list above is a hypothesis. Whether a new product page layout or a shorter checkout raised sales cannot be read from last month compared with this month, because season, campaigns and traffic mix move at the same time. An A/B test shows one version to half of your visitors and the current version to the other half, and compares the result in the same period.
This is how Varify.io fits into the four levers:
- Test each change. With Prompt-Based Experimentation (PBX) you describe the change in plain language. The AI builds the variation, writes the code and sets up the experiment, without a developer. Then compare the variation with the original. See PBX examples.
- Measure in your own GA4. The Tracking Setup Wizard connects the experiment to Google Analytics 4, and all GA4 events can serve as goals, so test numbers and analytics numbers match. See tracking with Varify.io.
- Use your shop data. With Shopify, revenue, purchase and order value can be imported as goals. See Shopify conversion data.
- No cookies. Varify stores the variation assignment in localStorage, without personal data. Read more in cookieless A/B testing.
Before you start, estimate how long a test needs. The free test runtime calculator uses your traffic and conversion rate, and the significance calculator checks a finished result. A test with little traffic can still be run, it simply needs a larger change to produce a clear answer.
Checklist: your first 30 days
Thirty days is enough to measure, choose one lever and run a first test. Adapt the order to the bottleneck you found in the first section:
Days 1–7: measure
- Write down visitors, conversion rate, average order value and repeat purchase share for the last full quarter.
- Open the funnel report in your analytics tool and note the stage where the most visitors leave.
- Check that your purchase event arrives in your analytics tool and matches the numbers in your shop system.
Days 8–14: choose
- Pick the one lever that looks weakest.
- Write three hypotheses in the form: “If we change X, more visitors will do Y, because Z.”
- Rank them by expected effect and by effort. Take the top one.
Days 15–30: test
- Describe the variation with PBX, let the AI build it, and check it on phone and desktop with the preview link.
- Estimate the runtime with the test runtime calculator and set the end date before you start.
- Start the test, leave it untouched until the planned end, and read the result against the goal you chose first.
- Roll out a winner, write down what you learned, and take the next hypothesis from the list.
Find out which change actually raises your sales.
Describe every change in plain language, let the AI build the variation and measure it in your own GA4.
Looking for more ideas to test? The conversion rate toolkit collects them in one place.