Insights.
How we think, what we fix first, and what we won't do. Opinion and method, never invented results.
Creative fatigue: how to see it coming
By the time an ad is obviously expensive, it has been wearing out for weeks. The signs arrive in order, and the early ones give you enough time to replace it calmly.
Reviewed 2026-09-08
Email list growth without pop-up fatigue
A list built by interrupting everyone on arrival behaves differently to one built at moments of real interest. The size looks similar. The revenue does not.
Reviewed 2026-09-08
GTIN, MPN and brand: the feed fields that decide visibility
Identifiers are how a shopping system decides your product is the same thing someone else is selling. Get them right and you join the comparison. Invent them and you leave it.
Reviewed 2026-09-08
How AI crawlers read a product page
A machine summarising your product cannot see your design. It reads what the page says, in order, and repeats the parts it can state confidently.
Reviewed 2026-09-08
How AI overviews changed product research
The research stage of a purchase is being answered above the results, which changes which pages are worth writing and how you measure them.
Reviewed 2026-09-08
How performance max spends your budget
A campaign type that decides its own placements still answers to the inputs you give it, and the spend pattern tells you exactly which input is wrong.
Reviewed 2026-09-08
How price competitiveness affects shopping ads
Shopping surfaces put your price next to the same product sold elsewhere. That changes what you can win on, and it is rarely solved by matching whoever is lowest.
Reviewed 2026-09-08
How returns quietly eat your ROAS
Every return is a sale the ad platform still counts. The gap between the ROAS on your dashboard and the money in your account shows up weeks later, and it is never spread evenly across campaigns.
Reviewed 2026-09-08
How seasonality hides revenue leaks
Strong seasons make weak pages look fine. The buyer arrives determined, forgives what would normally lose them, and the store gets credit for the calendar.
Reviewed 2026-09-08
How to brief an agency on your store
The quality of the advice you get back is set by the brief you send. Most briefs describe a mood and a wish. The useful ones describe an economy, a set of constraints and a history.
Reviewed 2026-09-08
How to handle out-of-stock pages without losing rankings
A product going out of stock is a merchandising event, not a reason to remove a page that took a year to rank. What the page should do while it waits.
Reviewed 2026-09-08
How to read a Shopify analytics dashboard
The default dashboard shows dozens of things and prioritises none of them, so the skill is knowing which four to read and in what order.
Reviewed 2026-09-08
How to read our monthly report
A report is only useful if the person reading it knows where to push. These are the checks we would apply to our own document, and the lines that should make you ask a question.
Reviewed 2026-09-08
How to set ad budgets by margin, not revenue
A budget set as a share of sales grows fastest on the products that can least afford it. Contribution margin gives you a ceiling you can actually defend.
Reviewed 2026-09-08
How to test a free shipping threshold
Moving the line is a five minute settings change. Knowing whether it worked takes a defined measure, a comparable period and the discipline to leave it alone while it runs.
Reviewed 2026-09-08
How we choose which lever first
Visitors, conversion, order value and repeat purchases each have an entry condition. Which one we open first is decided by dependency and evidence more than by which number looks worst.
Reviewed 2026-09-08
How we keep 700 pages honest
A site this size cannot be kept honest by good intentions. It needs rules that can be applied to every page and machinery that refuses the ones that fail.
Reviewed 2026-09-08
How we prioritise fixes across the four levers
Not the biggest idea first. The fix with the most money behind it, the most evidence under it, and the shortest wait before anyone can tell whether it worked.
Reviewed 2026-09-08
How we scope a quote
A quote should be readable backwards. If you cannot see which work produced the number, the number was produced by something else.
Reviewed 2026-09-08
Loyalty programs: when they pay and when they don't
Points and tiers are a discount you promise now and pay later. Whether the promise is worth making depends on purchase frequency, margin and what you would have earned without it.
Reviewed 2026-09-08
MER vs ROAS: which number to run the business on
Two ratios, two jobs. One tells you whether marketing is paying for itself, the other tells you which campaign deserves the next dollar.
Reviewed 2026-09-08
Mobile checkout: where the drop-off really is
Blaming the checkout is comfortable because it is a discrete thing you can rebuild, and it is usually not where a phone loses the order.
Reviewed 2026-09-08
Packaging as a repeat-purchase lever
Most stores treat packaging as a cost line or a branding exercise. It is closer to a channel, and it reaches the customer at the exact moment they decide whether ordering again is easy.
Reviewed 2026-09-08
Review schema: what it does and doesn't do
Markup is a translation layer, not a persuasion tool. It tells a machine what is already visible, and everything useful about it follows from that one idea.
Reviewed 2026-09-08
SMS consent done properly
A phone number collected carelessly is a liability and a poor performer. One collected properly is the most direct line a store has to a customer who wants to hear from it.
Reviewed 2026-09-08
Subscription churn: the first 90 days
A subscription is renewed or abandoned in the weeks around the second delivery, which is where nearly all the retention work belongs.
Reviewed 2026-09-08
The case against sitewide discounts
The discount you apply to everything is paid, in full, on the orders you were going to receive anyway, and the habit it teaches lasts longer than the sale.
Reviewed 2026-09-08
The checkout fields to remove first
Shorter checkouts convert better, but not every field is equally expensive. Some cost you orders, some cost you nothing, and a few are worth the friction they create.
Reviewed 2026-09-08
The difference between a leak and a plateau
Flat revenue has two very different causes, and the fix for one makes the other worse. Here is how to tell which one your store has before you spend against it.
Reviewed 2026-09-08
The first three fixes for most stores
Not because stores are the same, but because the cheap, high-certainty work tends to sit in the same three places: what the page fails to answer, what the checkout demands, and what happens after the order.
Reviewed 2026-09-08
The four numbers we report every month
A report exists to support a decision, so ours contains the four numbers that multiply into revenue and almost nothing else.
Reviewed 2026-09-08
The Merchant Center errors we see most
Nearly everything that takes products out of Google Shopping comes down to one thing: the feed says something your product page does not confirm.
Reviewed 2026-09-08
The month-to-month model explained
A twelve month agreement protects the supplier from a bad quarter. Month to month moves that risk onto us, and that changes what gets built first.
Reviewed 2026-09-08
The post-purchase email people actually open
Order confirmations and shipping notices get opened because someone is waiting on a parcel. Most stores fill that attention with a cross-sell nobody asked for.
Reviewed 2026-09-08
The product page questions that stop people buying
Most conversion work on a product page is not persuasion, it is closing the specific questions that a buyer will not check out with still open.
Reviewed 2026-09-08
The real cost of a free shipping threshold
Free shipping is not free, and the line you set decides whether it buys you bigger orders or just gives margin away to people who were already buying.
Reviewed 2026-09-08
The SEO cost of discontinuing products
Every product you stop selling takes an address with it. Done casually a few hundred times, that is how a store spends years running to stand still.
Reviewed 2026-09-08
UGC that sells vs UGC that just looks real
Handheld footage with nothing in it is still an ad with nothing in it. The value of a customer clip is the question it answers, not the fact that it was filmed on a phone.
Reviewed 2026-09-08
Wallets and express checkout: what to expect
Express checkout is one of the few conversion changes you can make in an afternoon. It is also one of the few that quietly changes your order value, your email list and your reporting.
Reviewed 2026-09-08
What a 1% conversion lift is worth at every revenue level
Conversion is the only lever that pays you on visitors you have already bought, which is why we usually look at it before we look at the ad budget.
Reviewed 2026-09-08
What a growth plan looks like for a $500k/month store
A store at this size is rarely broken. It is capped. The job of the plan is to name the cap, price the fix, and put the work in an order your team can actually staff.
Reviewed 2026-09-08
What a growth plan looks like for a $50k/month store
A store at this size has enough data to find its leak and not enough time to fix everything, so the plan is mostly about what to ignore.
Reviewed 2026-09-08
What apparel brands get wrong in product feeds
Apparel is the category where the feed carries the most information and where stores most often send a single row for something a shopper buys in twelve versions.
Reviewed 2026-09-08
What changed in Google shopping this year
Our reading of where control in Shopping actually sits now, and which of your own inputs decides what the system does with your money.
Reviewed 2026-09-08
What Google's shop tab means for small brands
When the storefront is assembled from your feed, the things you can control are your data, your images and your prices. That is a fairer fight than it sounds.
Reviewed 2026-09-08
What shoppers ask ChatGPT before they buy
The questions people put to an assistant are the ones they were too impatient or too embarrassed to research across ten tabs, and they are all answerable on your own pages.
Reviewed 2026-09-08
What the free growth plan is and isn't
Free usually means a template with your logo on it. Ours is four written pages about your store, and the most useful page is the one listing what we would not touch.
Reviewed 2026-09-08
What to expect in the first month with us
A month is long enough to fix something, measure it and decide what happens next. It is not long enough to hide behind a strategy document, which is why we do not write one.
Reviewed 2026-09-08
What we look at in the first hour on a new store
The first hour is spent as a buyer, not as an analyst. Shop the store, run the checkout, then open three reports. Anything else can wait until tomorrow.
Reviewed 2026-09-08
What we won't do for a store
Anyone can list what they offer. The more useful list is the one an agency turns down, because that is where the incentives are, and where you find out what happens when the work gets hard.
Reviewed 2026-09-08
When to turn off a campaign
Most campaigns are killed too early on a bad week or kept too long on a good story. The way out of both is a rule written before the campaign goes live.
Reviewed 2026-09-08
Why collection pages outrank product pages
People search for the category and buy the product. The page that matches the search is the one that keeps its position while your catalogue turns over.
Reviewed 2026-09-08
Why most abandoned cart emails are too early
An hour is the default delay in most email tools, and for most stores it interrupts a shopper who has not finished shopping yet.
Reviewed 2026-09-08
Why product titles matter more than ad copy in shopping
Search ads let you write the pitch. Shopping ads show your product title instead, so the title is both the thing that gets matched and the thing that gets read.
Reviewed 2026-09-08
Why ROAS is the wrong north star
A return figure can rise while the bank balance falls, which is why we treat it as a campaign dial rather than the number a store is run on.
Reviewed 2026-09-08
Why speed matters more on product pages than the homepage
Speed work follows attention, and attention inside a company points at the homepage. Your buyers arrive somewhere else entirely, usually on a phone, usually on the heaviest template you own.
Reviewed 2026-09-08
Why supplement brands get disapproved
In this category the review is reading your language, so the fix is almost always editorial rather than technical.
Reviewed 2026-09-08
Why the author on every page is the company
Content sites invent authors because a name and a face feel more trustworthy than an organisation. A byline nobody can check is decoration, and we would rather sign the work properly.
Reviewed 2026-09-08
Why we fix conversion before traffic
Conversion is the only lever that improves traffic you have already bought. Traffic only ever improves itself, and it charges you again every month.
Reviewed 2026-09-08
Why we never take a percentage of ad spend
Pricing decides which advice you get. A percentage of spend is a small tax on your ambition and a quiet reward for keeping the budget large, whatever the campaign is doing.
Reviewed 2026-09-08
Why your best customers never see your emails
Deliverability is judged on your whole list, so the subscribers who never open quietly decide whether your best customers ever see a thing.
Reviewed 2026-09-08
Your plan is free.
A person reads it. You keep the plan. No logins, no call first.