How to read your weekly numbers honestly

Introduction

Your first weekly report will probably be smaller than you expected, and your first instinct will probably be to decide what it means about you. Try to resist that. A week of numbers is not a grade. It is a short record of what happened on your site while you were busy doing other things, and it takes a few weeks of practice before it says anything you can trust.

WPMaven's operators work on a weekly rhythm. The team plans the week, does the work, and reports back what happened. That last part is where most new owners get stuck, because nobody ever taught them how to read it. The worry is usually the same one: what if I look at the wrong number, believe it, and build a bad decision on top of it.

That worry is reasonable, and the fix is boring. Learn what each number counts, learn what it stays silent about, and repeat one small routine every week. This guide covers both, and it assumes you have never done any of it before.

Decide what the week was supposed to prove

Numbers only mean something next to an intention. Before you look at anything, find the plan the week started with. Your Co-CEO decides what the week is for and tells you, so the honest first question is not "are the numbers good" but "did the thing we planned actually happen, and did it move what we expected it to move".

This matters more than it sounds. A week aimed at publishing content and a week aimed at fixing the checkout produce completely different results, and the two are not comparable. If you judge a content week by orders, you will conclude the work failed when it was never pointed at orders in the first place.

It also keeps you honest about your own part in it. Nothing goes live without your approval. If a batch of finished work sat waiting on you for four days, then the week's numbers describe a site that spent most of the week unchanged. That is not the work failing. It is a gap between what was ready and what was live, and you are the only person who can close it.

Write the week's intention down in one sentence somewhere you will see it again. When the report arrives, read it against that sentence instead of against your mood.

What visitors, orders, and revenue each actually measure

Three numbers get most of the attention, and they answer three different questions.

• Visitors tell you how many people arrived. That is the whole claim. Not whether they wanted what you sell, not whether they were the kind of person who buys, not whether they stayed longer than a second.
• Orders tell you how many people went all the way through and paid. This is the only number where someone made a decision that cost them something.
• Revenue tells you how much money came in through the checkout before anything is taken back out. It is orders multiplied by what people paid, adjusted for discounts, shipping, refunds, and taxes. It is not what you keep.

The useful reading is not any one of them alone. It is the steps between them. People arrive, some of them look at something you sell, some of them start the checkout, some of them finish. At every step, people leave. Comparing the steps tells you where the problem is, while comparing the totals only tells you that there is one.

Two habits make those steps readable. First, look at where the visitors came from, because people who followed a link you posted yourself behave nothing like people who found you through a search, and mixing them together hides both. Second, keep revenue and profit permanently separate in your head. Money that arrives is not money you keep, and the space between them is the part of the business your Chief Financial Officer is watching.

One more thing worth knowing early. A refund or a canceled order can change a week you have already read. The last few days of any period are the least settled, so give recent numbers a little room to move before you build an argument on them.

Why small numbers move so much

One easy mistake is treating a small count as a measurement. When few people are involved, ordinary randomness produces swings that look exactly like meaning.

If one more person buys this week than last week, the percentage change can look dramatic while nothing real has changed. If one fewer person buys, the same thing happens in the other direction and it feels much worse than it is. Percentages calculated on small counts are the least reliable thing on any report and also the most quoted. When a count is small, read the count and ignore the percentage.

Weeks are not interchangeable either. Traffic moves with the day of the week, with holidays, with paydays, with the weather in some trades and with the school calendar in others. One week compared against the week before it is a comparison of two noisy samples. Three or four weeks in a row, read as a line, tells you far more than any single jump.

Watch for weeks that were never really about customers. Launch week may be mostly friends, family, and your own testing. A post that got shared once can bring a crowd of people who were curious rather than shopping. The first week, the one that cost you a dollar, is a baseline and not a verdict. None of those weeks are lying to you, but none of them are evidence about demand.

The things your numbers cannot tell you

Numbers are a record of what people did. They contain nothing about why, and the why is what you need before you can decide anything.

Your report cannot tell you that the shipping cost surprised someone at the last step, that a photo made the product look smaller than it is, that a person meant to come back on payday, that someone recommended you to a colleague, or that your checkout is awkward with one thumb on a phone. Those are the real causes, and they live outside the count.

So when a number surprises you, treat it as a question rather than an answer. There are three cheap ways to get closer to the why, and none of them need any technical skill.

• Read what people actually wrote. Messages handled by your Customer Support Agent are the least filtered information you will ever get about your own business.
• Go through your own site on your phone as if you had never seen it, and buy something. Most confusing checkouts get discovered this way, not in a report.
• Ask. Your Co-CEO is the operator you talk to, so send the question there: which step lost people, where the traffic came from, what changed on the site that week.

Be careful with the opposite error too. A quiet week is not proof that nothing worked. A page published this week can be found by someone months from now. Some work has a delay built into it, and the week it shipped is the wrong place to look for its effect.

A weekly routine you can actually keep

Consistency beats depth here. A short review you repeat every week teaches you the shape of your own numbers, and that shape is what eventually lets you notice something real. Pick a day and a time, and keep it even on the weeks you are afraid of what you will find. Twenty minutes is enough.

1. Read the week's report first, before any raw numbers. It tells you what was planned and what was done, which is the context everything else needs.
2. Look at visitors, then the steps between arriving and ordering, then orders, then revenue, in that order. Working from the widest number inward stops you from panicking about a small number that was always going to be small.
3. Compare against the last three or four weeks, not against last week alone.
4. Write three lines: what changed, what we did, what I want to know.
5. Send that third line to your Co-CEO as a single question.
6. Clear your approvals. Work sitting in your queue is work that is not being measured.

Keep it all in the same file or notebook. After a couple of months, that boring list of what you thought each week becomes the most valuable document you own, because it shows you which of your instincts turned out to be right.

When to change something and when to wait

The honest answer to "should I change something" is that it depends. Here is what it depends on.

It depends on which step is failing. If almost nobody arrived, nothing downstream means anything yet, and the work belongs with your SEO Specialist, Social Media Manager, or Growth Concierge. If people arrived but never reached anything you sell, it is a navigation and messaging problem, which is your Web Designer and Content Writer. If people reached the checkout and stopped, look at shipping cost, payment options, and how much trust the page earns, which is your E-commerce Manager. Fixing the wrong step is the most common way to lose a month.

It depends on how many people were involved. With very few visitors, no change to a page can be judged, because the difference between one version and another is invisible underneath the noise. Below that point, the right move is almost always more traffic rather than more tweaking.

It depends on how long your buyers take to decide. An impulse purchase gets decided in the same visit. Something people think about for a week, or compare across three sites, will not show up in the week you changed it. Give a change at least as long as your customers take, plus a couple of weeks.

It depends on how many things changed at once. If the week included new photos, a new price, and a new post, you cannot credit or blame any one of them. That is fine early on when you want momentum, but once you are trying to learn something specific, change one thing and hold the rest still.

One rule applies no matter what. Do not approve work you cannot explain to yourself. You never have to know how it is built, but you should be able to say what it is for and what you expect it to move. If you cannot, ask your Co-CEO before you approve it rather than after.

Conclusion

None of this makes a bad week feel good. What it does is stop a bad week from turning into a bad decision, which is the more expensive of the two. Most weeks are not evidence of anything. A few are, and you will only recognize those if you have been reading the ordinary ones all along.

Start with the report you already have. Open it, write your three lines, and send one question to your Co-CEO about the step that lost the most people. Then do it again next week, at the same time, and let the pattern build.