How to read your own numbers

Your dashboard says 1,240 views this month. You made nine sales. You've spent twenty minutes staring at that and you have no idea whether it's good, bad, or the same as last month.

Four numbers are worth your attention at this size, and views isn't one of them. Orders. Average order value. Contribution per order. And how many people bought from you more than once.

That's it. Everything else, conversion rate, bounce rate, time on page, follower growth, is either noise at your volume or a number you can't act on. You need a few hundred orders before most of them mean anything, and until then they'll mostly convince you of things that aren't true.

Below is what each of the four is, how to work it out when half your sales happen in WhatsApp, and the honest problem with how long any of it gets kept.

Small numbers lie, and it's worth understanding why

If 40 people visit your shop in a week and one buys, your conversion rate is 2.5%. If two buy, it's 5%. Nothing changed about your shop. One extra person had a slightly better Tuesday.

That's the whole problem with reading rates at low volume. The swing from random variation is bigger than the swing from anything you'd do deliberately, so you end up attributing a normal fluctuation to the caption you changed on Thursday. Then you change more captions.

Rates get useful when the denominator is big and stable. Until then, count things instead of dividing them, because a count of nine orders is nine orders and it can't be misread. What conversion rate you should expect goes further into why the published figures for this are close to useless for a shop your size.

While we're on published figures: you'll find a lot of confident industry averages for conversion rate, abandonment, repeat purchase and email open rates. Most of them are unsourced, or they trace to a study of large stores in one country in one year, and they get copied between blogs until they look like facts. Don't benchmark yourself against them. Benchmark this month against your own last three months, which is the only comparison that describes your shop.

The four numbers

Orders

The count of orders in the month. Not revenue, not views. Orders.

It's the least sophisticated number available and it's the one that tells you most, because it goes up when things are working and down when they aren't, and there's nothing to misinterpret.

Count every order regardless of how it arrived. A WhatsApp order is an order. A cash-on-delivery order that was delivered is an order. One that was refused at the door isn't.

Average order value

Total revenue divided by number of orders.

This one moves for reasons you control: bundles, a higher-priced hero product, free delivery over a threshold, or simply raising prices. If it's drifting down without you deciding to make it drift down, something in your catalogue is pulling people toward the cheap end.

Contribution per order

Revenue minus the costs that vary with the order, divided by orders. Materials, packaging, payment fees, and whatever share of delivery you absorb.

This is the number that tells you whether more orders is actually good news. A shop doing 40 orders a month at $4 of contribution each is doing worse than one doing 15 at $19. Plenty of small sellers grow their order count and get poorer, and they don't find out for a year.

Full profitability, once you add rent, subscriptions and your own wages, is a bigger calculation and it belongs in a quarterly sit-down rather than a monthly glance. Contribution per order is the fast version, and it's the one that catches the problem early.

Repeat rate

Of the people who bought this month, how many had bought from you before. Expressed as a count, not a percentage, until you're doing enough volume for a percentage to be stable.

"Three of this month's eleven buyers had bought before" is a sentence with real information in it. It's also the number that predicts next year, and it's the last one anybody checks.

A shop where nobody comes back is a shop that has to find every customer from scratch, forever. That's an exhausting business and it's usually a product or a follow-up problem rather than a marketing one.

The fifth number, if you're spending money

If you pay for anything to get customers, ads, a market stall, boosted posts, sponsored content, add one more: cost per order.

Total spent divided by orders you can attribute to it. Attribution is fuzzy at small scale, so use a rough method and be consistent: ask new customers where they came from, and count the ones who say the paid thing.

Then compare it to contribution per order. If you're spending $14 to acquire an order that contributes $9, stop. That's not a growth phase, it's a leak.

The record that takes two minutes a month

One row per month, filled in on the first of the following month. In a spreadsheet, not in your head.

Month	Orders	Revenue	AOV	Contribution	Repeat buyers	Notes
Jan	11	$412	$37	$198	3	Sale on the 14th
Feb	9	$306	$34	$151	2	Nothing unusual
Mar	14	$488	$35	$241	4	Reshot the mug listing

The notes column is the part that makes the rest useful. Twelve months from now, a dip in February means nothing without a note saying you were away for ten days.

Fill it in on the first. Before you've decided how the month felt. A number recorded after you've formed an opinion about the month has a way of coming out matching the opinion.

Where the numbers come from when your sales live in WhatsApp

Most small sellers have orders scattered across a shop link, DMs, comments and messages from a group chat, which means no single dashboard sees all of it.

Two options, and the simple one is better.

Option one: one place where every order gets written down. A sheet, four columns: date, customer, amount, channel. Add a row when the money lands. Ten seconds per order, and it is the only view of your business that's actually complete.

Option two: reconcile monthly. Take your shop analytics, add the orders that happened only in DMs, and total them on the first. Faster per order, more likely to miss things, and it falls apart if you're busy.

Whichever you pick, the channel column earns its place fast. Within three months you'll know whether Instagram or the estate group or the shop link is where orders actually come from, and it's regularly not the one that gets most of your effort.

Alina's mugs, in Portland

Alina throws ceramic mugs in Portland. $34 each, $46 for a pair, $7 shipping in the US, and she sells through a shop link she posts to Instagram.

Her first year she watched views. Some months 900, some months 2,400, and she'd feel good or bad accordingly. She could not have told you her average order value.

She started the monthly sheet in September. Six months of it:

Month	Orders	Revenue	AOV	Contribution	Repeat
Sep	12	$421	$35	$198	1
Oct	15	$532	$35	$250	2
Nov	22	$859	$39	$412	3
Dec	31	$1,284	$41	$611	5
Jan	8	$272	$34	$126	2
Feb	11	$398	$36	$187	4

Three things came out of that sheet which no view count would have shown her.

Average order value rose from $35 to $41 across the autumn, because she'd started offering the pair at $46 and roughly a third of buyers took it. That's $6 an order for a change that cost her nothing but a second listing.

January's collapse looked like a disaster until she put it next to December. It's a season, and knowing that in year two meant she didn't panic and discount, which is exactly what the quiet months argues for.

And repeat buyers went from 1 to 4 a month over six months. Four out of eleven in February. That's the number she now cares most about, and it arrived slowly enough that she'd never have noticed it without the sheet.

The month that taught her the most was March, which isn't in the table. Her traffic doubled after a post did unusually well, and her orders didn't move at all. That told her the problem was the page rather than the posting, and it's the only month in two years where a view count was worth reading. A number is worth looking at when it moves independently of the others, and not otherwise.

Telling a real change from noise

Three rules that will save you a lot of bad decisions.

One month is not a trend. Two consecutive months in the same direction is worth noticing. Three is a trend. If you react to every month, you'll be changing your entire shop every four weeks and you'll never know what caused anything.

A change under about 20% at low volume is probably noise. Nine orders to eleven is not growth. Nine to eighteen is.

Change one thing at a time. If you reshoot the photos, raise the price and start posting more in the same week, you've learned nothing from whatever happens next. This is boring advice and it's the difference between a seller who improves and one who thrashes.

When to act on each number

The number	If it's falling	If it's rising
Orders	Check enquiries first. Fewer enquiries is reach, same enquiries is conversion	Work out which channel, and do more of that
Average order value	Something's steering people to the cheap option	Keep whatever you changed, and push it further
Contribution per order	Costs moved, or you discounted more than you meant to	Consider whether you could raise price again
Repeat buyers	Follow-up has stopped, or the product disappoints on use	This is the good one. Protect it

The most common misread is falling orders with steady enquiries. That's not a marketing problem, it's something between the question and the payment, and finding out why people abandon is where to go for it.

The retention limit, plainly

Here's the constraint worth knowing before you rely on any dashboard, including Sailo's.

Sailo keeps 7 days of analytics on the free plan, one year on Pro at $19 a month, and three years on Business at $49 a month.

7 days means you can see whether this week is working, and nothing else. You cannot hold it against the week before, let alone this December against last December. You cannot see whether your average order value has drifted over eight months. For seasonality, which is most of what a small shop needs from analytics, 7 days is the wrong tool, and no amount of checking it more often fixes that.

A year on Pro is the minimum that lets you do a year-on-year comparison at all, and it's the main reason a seller with a seasonal business ends up on Pro rather than because of the product limit.

Even three years on Business runs out. If you're still trading in year four, the first year is gone.

Which is why the two-minute monthly sheet isn't a workaround, it's the actual answer. Your spreadsheet outlives every dashboard you'll ever use, it captures the DM orders no analytics tool can see, and it's portable if you ever change platform. Use the dashboard for the detail inside a month. Use your own record for the shape of your years.

Fill in one row tonight

Open a spreadsheet. Make the seven columns: month, orders, revenue, average order value, contribution, repeat buyers, notes.

Fill in last month from whatever records you've got. It'll take longer than two minutes this once, because you're reconstructing. Then set a recurring reminder for the first of every month.

By month three you'll be able to answer the question you couldn't answer at the top of this article. And once you can see contribution per product rather than per order, the next decision is which one to put in front of people, which is how to know which product to push. If you're not yet at a point where there are numbers to record, getting your first ten orders comes first, and the sheet starts on order one.