We do not start in the analytics. Analytics tells you where people stopped, and it is very good at that, but it never tells you why, and the why is what a first hour is for.
So the first hour on a new store is spent as a buyer. Phone in hand, on data rather than office wifi, the way most of your customers arrive. Find a product the way they would find it, read what the page tells you, put it in the cart, and go all the way to the payment step. Most of what we end up recommending in month one is visible in that half hour, before anyone has opened a dashboard.
The order matters more than the list. Experience first, because it generates the questions. Data second, because it tells you how much each question is worth. Doing it the other way round produces a tidy report about a store nobody has actually used.
Ten minutes shopping it on a phone
Not browsing it. Shopping it, with the intention of buying something specific, on a phone, with the sound on.
We are looking for the moment we have to work. Where the page makes us guess a size, or scroll past four blocks of brand copy to find what the thing is made of, or open a tab to check whether it ships to us. Every one of those moments is a place where a real buyer with less patience and no professional obligation simply leaves.
A few things get noted almost every time. How long the product page takes to become usable, not to finish loading, but to be scrollable and tappable without things jumping. Whether the price and the delivery cost appear anywhere near each other. Whether the photos answer scale, so you can tell how big the thing actually is. Whether the reviews are about this product or about the store in general.
We also note what is good. Stores usually have one thing they do unusually well, and it is often buried below three things that were added later. Part of the first hour is finding that and asking why it is not the first thing a buyer meets.
We start on a phone with a card in hand, not in an analytics tab.
The checkout, all the way to the payment step
We take a real cart to the last screen before payment on every store, because that is where the surprises live and no one who works on the store sees it with fresh eyes any more.
The questions are simple. When does shipping cost first appear, and is it a number or a promise to tell you later. How many fields are there, and how many of them could be filled from a wallet or an address lookup. Is a customer account required, or offered. Are the payment methods the ones your buyers use, in the order they expect. Does anything ask for information the store does not need, like a company name field on a consumer store, or a phone number with no explanation of what it is for.
Then we do it again on a slow connection, and again from a different country if the store sells across borders, because currency, duties and delivery estimates all change the answer.
This is unglamorous work and it consistently finds money. A checkout that adds a step, a form that rejects a valid postcode format, a delivery estimate that only appears after payment details. None of it shows up as a broken page. All of it shows up as a number.
Then three reports, in this order
Now the data, and only three things to begin with. More than three in the first hour and you start pattern matching on noise.
First, sessions and orders by source, over twelve months. This says where the business actually comes from and how concentrated it is. A store with nine tenths of its orders from one source has a different first problem to a store with a spread, no matter what the totals look like.
Second, conversion rate by device. Nearly every store converts worse on mobile than desktop, so the question is not whether there is a gap but how wide it is against the share of traffic on each. A wide gap on a store where most visits are mobile is usually the largest number available in the whole account.
Third, orders by product, with the long tail visible. This shows concentration, it shows which products can carry ad spend, and it shows the products that get traffic and no orders, which are often the most interesting rows on the page.
That is enough to size the questions the first half hour raised. Everything else, the cohort work, the flow reporting, the auction data, comes later in the week when we know which of them is worth the time.
Your own products, searched the way a buyer searches
The last part of the hour is spent outside the store, looking for it.
We search a few of the store's real products the way a buyer would type them, including the plural category phrase, and see what comes back. Does the store appear at all. Is it the product page, a collection page, a marketplace listing of the same product, or a reseller. Does the shopping surface show the product, with the right price and the right availability, and does the image hold up next to the others in the row.
This is where feed problems announce themselves before anyone opens the feed. A product that is in stock on the store and shown as unavailable elsewhere, a price that does not match, a title that starts with an internal code, a product missing entirely while its variants are present. Each of those is a specific, checkable fault, and each is usually costing traffic that has already been paid for in some other way.
We also search the brand name on its own. What people find when they look you up directly is the cheapest impression the store gets, and it is often the one nobody has looked at in two years.
What we write down, and what we ignore for now
At the end of the hour there is a page of notes, and it is deliberately not a to do list.
What goes on it: the exact moments where buying was harder than it needed to be, with the step named. The three reports, with the numbers copied down and dated, so there is a baseline before anything changes. Anything factually wrong, like a price mismatch or a broken link in the checkout path. And the two or three questions the hour raised that the data has not answered yet.
What we ignore, on purpose: design taste. Whether we like the fonts is irrelevant and saying so wastes the store owner's attention. Also ignored on day one, anything that requires a rebuild, any channel the store does not currently run, and any recommendation we cannot connect to one of the four numbers.
The hour is not the plan. It is what makes the plan specific rather than generic, and it is the difference between a document that could describe any store and one that could only describe yours. If a recommendation could be copied onto another store without changing a word, we have not done the hour properly.