How to estimate any store's sales

No outside tool can see a store's checkout, so every number you see for a competitor's sales is a model. That does not make the numbers useless. A clear, simple model applied the same way to every store is very good at telling you which stores are bigger, which are growing and roughly what size of business you are looking at. This guide explains the model RivalSonar uses, step by step, so you can check it, use it by hand or adjust it with your own knowledge.

Key takeaways

• Sales estimates come from three parts: monthly visits, a conversion rate and an average order value.
• Visits are modelled from public traffic rankings using a rank-to-visits curve.
• The conversion rate starts at about 2% and is adjusted for price band and category, then kept between 0.3% and 5%.
• Average order value is set at about 1.2 times the store's typical product price.
• The result is an estimate with a real margin of error. Use it to compare stores and spot trends, not as an exact figure.

The formula

Estimated monthly revenue = estimated visits x estimated conversion rate x estimated average order value

Estimated monthly orders are the first two parts on their own, visits times conversion rate. Each part has its own uncertainty, so it helps to look at them one by one.

Part 1. Estimating monthly visits

Several public sources rank websites by popularity. They publish a rank, not a visit count. Rank 1 gets vastly more traffic than rank 10,000, and the drop-off follows a fairly smooth curve. We fit that curve, a rank-to-visits curve, using sites whose traffic is known or widely reported, then read each store's estimated monthly visits off the curve from its rank.

Three things to know about this step:

• It is relative. A store ranked higher almost always gets more traffic than one ranked lower, even if the absolute numbers are off.
• It is noisier for small stores. Rankings are less stable in the long tail, so estimated visits for small stores move around more.
• It covers all traffic. The ranking does not care whether visitors come from ads, search or email. The traffic sources breakdown is a separate estimate.

If you have your own traffic data for a few stores, for example from partners or your own sites, you can check how the curve fits in your niche.

Part 2. Estimating the conversion rate

Ecommerce conversion rates, the share of visits that end in an order, cluster around 1% to 3% for most stores. We start from a base of about 2% and adjust it:

• Price band. Cheaper products convert better. A store whose typical product costs under about 25 gets nudged up. A store selling items in the hundreds gets nudged down, since people browse longer and buy less often.
• Category. Some categories convert better than others. Consumables and accessories tend to convert higher than furniture or high-end electronics.

The adjusted rate is then clamped between 0.3% and 5%. No store gets an estimated rate outside that range, because rates beyond it are rare and usually signal bad input data rather than a real outlier.

Part 3. Estimating average order value

We take the store's typical product price, the middle of its catalogue prices weighted toward its best sellers, and multiply it by about 1.2. The 1.2 reflects that many orders contain more than one item, add-ons or an upsell. Stores that push bundles or quantity discounts may run higher, and single-product stores with no add-ons may run lower.

A worked example

Here is the model applied to an imaginary store, with round numbers to keep the arithmetic easy.

You would read this as "an estimated 1,300 orders and an estimated 48,000 in revenue a month", not as a precise number. If the same store showed an estimated 2,000 orders the month before, the direction (down) is more trustworthy than either figure.

Splitting orders across products

A store's estimated orders do not tell you which products sold. RivalSonar splits them across the store's best sellers with a power curve, so the top product in the best-seller rank takes the biggest share, the second takes less, and the long tail shares what is left. This matches how most stores look from the inside, where a few products carry most of the sales.

How tracking improves an estimate

For stores you add to the Sales tracker, we refine the estimate hourly by watching catalogue changes, such as stock levels where they are public and changes in best-seller order. The first report is ready after 24 hours. These signals are still indirect, so tracked figures remain estimates, but they react to real changes much faster than a monthly model. See how to track competitor sales daily.

Where estimates go wrong

Know the main error sources so you can adjust in your head:

• Traffic that does not buy. Stores with viral content, blogs or giveaways get lots of visits that never convert, so their sales estimates run high.
• Unusually strong conversion. Stores with loyal repeat buyers, email-heavy marketing or very strong offers can convert well above 5%, so their estimates run low.
• Multiple storefronts. A brand that sells through several country domains or marketplaces may look smaller than it is, because each domain is estimated on its own.
• Wholesale and offline sales. Orders placed outside the online store are invisible.
• Big recent changes. Rankings lag. A store that just launched a huge campaign may not show it for a while.

Building your own estimate by hand

If you want to check a number yourself:

1. Find the store's rank in a public traffic ranking and convert it to visits with your own reference points.
2. Pick a conversion rate between 1% and 3% based on price and category.
3. Look at the product feed or collections to find a typical price, and multiply it by about 1.2.
4. Multiply the three. Then ask whether the result fits what you see in their ads, reviews and catalogue changes.

If the store has 40 reviews in total and shows an estimated 5,000 orders a month, something is off. Use your judgement and the other signals described in how to research a competitor's Shopify store.