Traffic, conversion, or order value — what is actually holding my store back?
Last updated: 2 August 2026
The factor holding you back is the one furthest below what a store like yours normally does — not the one that feels worst, and not all three. Revenue is sessions × conversion rate × average order value, so score each factor against a benchmark, take the widest gap, and work on that one until it stops being the widest.
Score the three factors
Pull 90 days from Shopify Analytics and calculate each one.
| Factor | How to calculate | Typical range | What a gap here looks like | | --- | --- | --- | --- | | Sessions | Straight from Analytics | — (compare to your own trend) | Everything downstream is healthy, there is just not enough of it | | Conversion rate | Orders ÷ sessions × 100 | ~1.5–3% for most stores | Plenty of visitors, few orders | | Average order value | Sales ÷ orders | Category-dependent | Orders arrive steadily, each one is small |
Then two more that are easy to forget:
| Factor | How to calculate | Why it matters | | --- | --- | --- | | Repeat rate | Returning-customer orders ÷ all orders | Cheapest revenue you will ever get | | Gross margin | (Sales − cost of goods) ÷ sales | A revenue win at a bad margin is not a win |
Pick the widest gap, not the worst feeling
The rule that saves the most wasted work: fix traffic last if conversion is broken. Sending more sessions into a funnel converting at 0.8% buys you more of the same problem at a higher cost. Get conversion into a normal range first, then the same traffic is worth roughly twice as much.
The reverse also holds. If your conversion rate is 2.6% and AOV is healthy, you do not have a conversion problem — you have a traffic problem, and another round of button-colour tests will not find it.
The check most people skip
Compute orders ÷ sessions before you trust your conversion rate. If it comes out above 100%, your conversion rate is not a conversion rate: Shopify counts every order (POS, draft, phone, wholesale) while sessions counts online-store visits only. The ratio is measuring two different populations, and a suspiciously high number will make a leaking funnel look healthy.
The honest move is to treat conversion as unreadable until the two series match, rather than acting on a figure that cannot be what it claims.
Why one factor at a time
If you change traffic, conversion and AOV in the same month and revenue moves, you have learned nothing about which change did it — so you cannot repeat it. One factor per cycle is slower per experiment and much faster per lesson.
Who this is not for
- Stores with fewer than about 30 days of sales. The percentages will swing on single orders.
- Stores where sessions are not tracked. Without sessions, conversion cannot be calculated at all, and a traffic diagnosis is guesswork. Fix the tracking first.
- Anyone who already knows the answer. If a specific thing is obviously broken — checkout erroring, a channel that stopped — go fix it. This is for when nothing is obviously wrong and revenue is flat anyway.
Common questions
What if two factors are equally bad? Take the one that is cheaper to move. Conversion and AOV changes are usually work you already control; traffic often costs money.
How long before I know it worked? One cycle of four weeks, minimum. Weekly numbers are too noisy to read as a trend.
Is margin ever the constraint? Yes, and it is the one most often missed — but only if you have cost-per-item on enough of your catalogue to calculate it. If you do not, that is the first job, not a margin plan.
Paceloop scores these factors against benchmarks on your own 90 days, names one constraint, and says which levers it ruled out and why. It refuses to name one at all when the data cannot support it. Install free — the diagnosis is part of the free audit.