GuideSix levers, and popups are the fifth.
This is written by a company that sells popups. Putting them fifth is not modesty — it is where they actually sit, and a list that put them first would be worth nothing to you.
In order of how much they move.
The first two are unglamorous and usually the largest.
Fix the checkout
Forced account creation, surprise shipping costs revealed at the last step, and a form asking for more than it needs are the three most reliable ways to lose a buyer who had already decided. This is not glamorous and it is almost always the biggest single number available.
Start here even if you are certain it is fine. Especially then.
Make the site fast on a phone
Most ecommerce traffic is mobile and most mobile traffic is impatient. Speed is not a conversion tactic so much as the condition under which every other tactic gets a chance to work.
Measure on a mid-range device on a real connection, not on your laptop.
Answer the question that stops the purchase
Sizing, materials, delivery time, returns. Every category has one or two questions that reliably halt a decision, and the answer is usually available somewhere on the site but not where the decision is being made.
Your support inbox already contains this list. Read a week of it.
Show the product properly
More images, on a real person or in a real room, at the scale the buyer needs. This is the least technical item here and frequently outperforms months of testing button colours.
Cheaper than most agencies will tell you.
Capture the visitors who will not buy today
The large majority of your traffic leaves without buying, and most of them were never going to on this visit. Capturing an address converts a lost session into a second chance, which does not raise today's conversion rate but raises the one that matters over a quarter.
This is where a popup belongs — fifth, and honestly described.
Test one thing at a time
Most stores plateau not because they run out of ideas but because they changed four things at once and cannot attribute the result. A slower test you can learn from beats a fast one you cannot.
And leave it running long enough to mean anything.
What lever five is actually worth.
Capturing an address does not make today’s visitor buy today. It converts a session that was going to end in nothing into a chance to reach that person again, and the return shows up in revenue over weeks rather than in your session-to-order rate this afternoon. That is a smaller and slower claim than most popup marketing makes, and it is the one we can actually stand behind.
Conversion rate, answered plainly.
Including the benchmark question, which has no honest single answer.
We are not going to publish a number, and you should distrust pages that do without saying where it came from. Rates vary by an order of magnitude across categories, price points and traffic sources — a store selling £8 consumables to returning customers and one selling £2,000 furniture to cold paid traffic have nothing meaningful in common. Your own rate last quarter, split by device and by traffic source, is the only benchmark that will tell you anything actionable.
Usually removing something from the checkout: a forced account, a required field, or a shipping cost that appears only at the final step. These are cheap to change, they affect buyers who had already decided, and the effect shows up quickly. Most stores reach for acquisition tactics first and audit the checkout last, which is the wrong order.
Not the one most people mean. A popup does not usually make more of today's visitors buy today — it captures the ones who are leaving without buying, so you can reach them again. That is genuinely valuable and it shows up in revenue over weeks, not in your session-to-order rate this afternoon. Anyone selling a popup as a direct conversion-rate lever is describing a different metric from the one you are measuring.
Long enough to cover the full cycle of your traffic — including weekends, paydays, and any campaign that skews a particular day — and long enough to accumulate a sample you would defend. Most ecommerce tests are stopped early, on a result that looked decisive and was noise. If the honest answer is that your traffic cannot support a test in reasonable time, that is a real finding: make the obviously-better change and skip the test.
Neither in isolation, which is why both can be gamed. A deep sitewide discount raises conversion rate and destroys margin; an aggressive upsell raises order value and costs completions. Revenue per session is the measure that resists both distortions, and it is usually available in the same analytics you already have.
As a second-chance mechanism rather than a first-visit one. The majority of ecommerce traffic leaves without buying and a large share was never going to buy on that visit. Capturing an address turns some of those sessions into a future one. It sits below checkout, speed and product presentation in impact, and above almost everything else — which is roughly fifth on any honest list.