What conversion rate should you expect

Somebody told you a good conversion rate is 2%. You worked yours out, got 0.9%, and now you're either depressed or about to rebuild your shop.

Don't. That 2% almost certainly came from nowhere.

The short answer is that nobody can tell you what your rate should be, and most of the figures you'll find were copied from an article that copied an article. There is one benchmark I could actually trace to a stated sample, and it doesn't describe a shop like yours. What follows is that number with its provenance attached, why it won't help you much, and the two things you should measure instead, which will.

The one number I could source, with its caveats attached

Littledata publishes a Shopify conversion benchmark. Reading their own page in August 2026, it says they benchmarked 2,800 Shopify sites in 2023 and found an average conversion rate of 1.4%, with the top 20% of stores above 3.2% and the top 10% above 4.7%. They split it further: mobile averaged 1.2% against desktop's 1.9%, fashion averaged 1.9%, food and beverage 1.5%.

That's a real figure with a stated sample and a stated year. It's the only one in this space I found that told me both.

Now the reasons it isn't about you.

It's Shopify stores, which means shops with a hosted checkout, a theme, usually a domain, and often a paid app stack. If your shop is a link from a bio and half your orders complete in a WhatsApp thread, that's a different animal.

It's session-based, so a person who visits three times and buys once counts as one conversion in three sessions. Visitor-based measurements produce a materially higher number from identical behaviour. That single choice of denominator explains a lot of the spread between sources.

It's from 2023, which the page states clearly.

And the average sits below the low end of most ranges you'll see quoted elsewhere, which should tell you something about how those other ranges were assembled.

Why the other numbers you'll find are not evidence

Go and search for a 2026 conversion rate benchmark and you'll get a page of articles with figures between roughly 1.4% and 3%. Read ten of them and you'll notice something: they cite each other, or they cite a vendor blog that cites another vendor blog, and the trail either loops or stops at a page with no methodology on it at all.

Some of them are honest and disclose their sample. Most quote a number with no denominator, no year, no sample size, and no statement of whether it counts sessions or people.

A percentage without a denominator is not a statistic. It's a decoration.

That's not a conspiracy either. It's what happens when a topic gets covered by a few hundred people writing for search rather than for accuracy, and the incentive is to have a number rather than to have a defensible one.

So the practical advice is: treat any conversion figure without a stated sample and year as though it were made up, because operationally it is. And treat your own number, measured consistently, as the only benchmark that can tell you anything.

Before anything else, define your denominator

This is where most sellers get a wrong number without realising, and it's the thing that makes cross-shop comparison close to useless.

Conversion rate is orders divided by something. The something can be:

• Sessions. Every visit, including the same person three times. Lowest number.
• Unique visitors. Deduplicated people. Higher, sometimes by half again.
• People who opened a product page, rather than the shop front. Much higher, because you've filtered out the browsers.
• People who started an order. Higher still, and this is checkout completion rather than conversion.
• People who messaged you. Very high, and a different measurement altogether.

Two shops with identical behaviour can report 1.2% and 6% depending purely on which of those they picked. When someone tells you their rate, and they haven't told you the denominator, they've told you nothing.

Pick one, write it down, and never change it, because the only comparison that will ever be useful to you is your own number against your own number last month.

Your number is noisier than you think

Here's the part nobody puts in these articles, and it's the reason most small sellers draw conclusions they shouldn't.

If you've had 200 visits and 4 orders, your rate is 2.0%. The range that number is actually consistent with, using a standard confidence interval for a proportion, is roughly 0.8% to 5.0%. Your true rate could be less than half what you measured, or more than double it, and 200 visits cannot tell the difference.

Orders / visits	Measured rate	Range it's consistent with
2 / 100	2.0%	0.6% to 7.0%
4 / 200	2.0%	0.8% to 5.0%
10 / 500	2.0%	1.1% to 3.6%
20 / 1,000	2.0%	1.3% to 3.1%
25 / 1,000	2.5%	1.7% to 3.7%

Read the first two rows again. At the traffic most small shops actually have, the number you calculate is nearly meaningless on its own. It certainly cannot tell you whether the change you made last week helped, because the change would have to be enormous to show above that noise.

The consequence is practical, not academic. Do not A/B test at this size. Do not change your prices because the rate dipped from 2.4% to 1.8%. Do not conclude your new photos worked because it went up. At a few hundred visits a month, a run of good luck and a genuine improvement look identical, and you'll spend months chasing your own noise.

What you can conclude at small numbers: zero orders from 400 visits is a real signal. Going from 1–3% to 4% is a real signal. Anything in between needs more data or a different question, and what to do when nobody is buying is the diagnostic that doesn't require statistics.

The average of your traffic describes none of your traffic

This is the most useful thing in the article and it's why a single shop-wide number misleads.

Anjali sells handmade banana-fibre totes from Kochi. They're ₹950, delivery inside Kerala is ₹90, and she takes payment by UPI to her own UPI ID, with about a third of orders going cash on delivery.

Over three months: 1,840 shop opens, 47 orders, 2.55%. Broadly in line with what she'd read, so she assumed she was fine.

Split by where the traffic came from:

Source	Opens	Orders	Rate
Instagram bio link	980	14	1.4%
Forwarded WhatsApp links	260	24	9.2%
One reel that travelled	600	9	1.5%
All	1,840	47	2.55%

The 2.55% describes none of her traffic. It's a blend of a 9.2% channel and two channels under 1.6%, and the blend moves entirely with the mix rather than with anything she does.

That table changed how she spent her week. Forwarded links were 14% of her traffic and 51% of her orders. Every hour she'd been putting into reels was buying her the worst-converting visitors she had, and the thing that actually worked, people sending her link to a friend, she'd never once deliberately asked for. She started asking, which is what asking customers for referrals is about, and her overall rate went up without a single change to the shop.

One more thing that came out of the split. Of her 47 orders, three were cash on delivery and refused at the door, so 44 completed. Her conversion rate as measured counted orders placed, not money received, which is a real distinction on manual payment rails and one that flatters every seller who doesn't check.

Your number is a mobile number, so compare it to mobile ones

The Littledata split is the most useful part of that benchmark, and it's the part nobody quotes. Mobile averaged 1.2% against desktop's 1.9%, and the gap widened at the top: their top 10% of stores hit 3.9% on mobile and 6.5% on desktop.

Now consider where your traffic comes from. A bio link, a forwarded WhatsApp message, a story swipe-up, a QR code on a market stall. Essentially all of it is a thumb on a phone, often on mobile data, frequently while the person is doing something else.

So if you're going to compare yourself to anything, compare yourself to the mobile figure, not the blended one. Sellers who benchmark against a desktop-inflated average conclude they're failing when they're roughly normal for their traffic.

The practical consequence is about the page rather than the number. Assume one thumb, one hand, and a screen about 390 pixels wide. That means the price visible without scrolling, buttons big enough to hit while walking, no pinch-to-zoom to read a size chart, and photos small enough to appear before the person gives up. Test it on a phone with wifi off, which is the single check most sellers have never done on their own shop.

For a lot of small sellers, conversion isn't the right metric at all

If your orders mostly complete in a chat thread, "conversion rate" is measuring a step your buyers don't take.

Split it into two numbers instead:

Opens to messages. Of the people who open your shop or your link, how many message you? Anjali got 118 messages from 1,840 opens, so 6.4%. This measures whether your page makes people want the thing and whether the price is doing its job.

Messages to orders. Of those 118, 47 became orders, so 40%. This measures your replies: your speed, your wording, your payment options, whether you answered "how much" with a full stop or with a next step.

Those two numbers point at completely different problems and the blended rate hides both. A shop at 2% could be 10% of openers messaging and 20% of those converting, which is a closing problem, or 3% messaging and 66% converting, which is a page problem. Same rate, opposite fixes, and you cannot tell them apart from one percentage.

Count messages for a month. It's a tally in a notebook and it will tell you more than any dashboard.

What actually moves the number

For small shops, in rough order of effect, and none of these are subtle:

A visible price. "DM for price" costs you most of the people who would have bought.

Delivery cost shown before the last screen. A cost that appears at the end reads as a trick, and it's where a lot of small-basket orders die. Put it in the listing next to the price, or set one flat number and say it everywhere.

Load time on mobile data. Not on your wifi. Photos straight off a phone camera are around ten times larger than they need to be, and people on a patchy connection don't wait.

Payment methods your buyers already use. A stranger who expects cash on delivery and finds only card will leave, and the reverse is equally true. Match the market.

Fewer products on the front page. Twenty items is a decision. Five is a purchase.

Proof somebody else bought it. One specific customer sentence under the price outperforms a page of five-star ratings.

Every one of those is a fix you can make this week, and each of them affects the rate more than anything you'd learn from a benchmark.

Where the tool's limits sit

Two honest ones, and both matter specifically for this question.

On the free plan you get 7 days of analytics. Pro gives you a year, Business three. If you're on free and you want to know whether this August is better than last August, you can't, because the data isn't there to look at. Keep your own monthly tally in a spreadsheet from day one and you'll never be stuck behind that, on this tool or any other.

The second is more subtle. On manual rails, Sailo cannot tell you the money arrived. Bank transfer, cash on delivery, a UPI payment straight to your own ID: the shop records that an order was placed, and only you and your bank know whether it was paid. So any conversion number you get counts intent, not settlement, and Anjali's three refused deliveries sit inside her 47 looking exactly like a sale. Whatever you measure, keep a separate count of orders that actually turned into money, because that's the one your rent is paid from.

Do this instead of looking up a benchmark

Three columns, one row a month, in whatever you already use.

1. Opens. Whatever your denominator is. Write down which one you chose.
2. Messages. A tally.
3. Orders that were actually paid. Not orders placed.

Fill it in on the last day of every month. After three months you'll have your own baseline, which is the only benchmark that can tell you anything, and after six you'll be able to see a seasonal shape.

Then split the first column by where the traffic came from, even roughly. That split is where the decisions are, and it's why how to read your own numbers is worth more than any industry average you'll ever be quoted.

And if you were about to spend money on traffic because the rate looked acceptable, do the split first. A 2.5% blended rate made of one great channel and two bad ones means paid traffic will land in the bad ones, which is exactly the trap in when to spend money on ads. If your traffic is still measured in the low hundreds, none of this applies yet, and getting your first ten orders is the job in front of you.