Pinterest, properly
Pinterest is a search engine that people have mistaken for a mood board, which is why it is the one platform where the work you do this month is still selling for you the month after next. Four hundred and fifty million people arrive typing what they want rather than scrolling to be entertained, most of them planning something they have not bought yet, and almost nobody is competing for those searches properly.
Every other platform pays you once. You post, it goes out, it dies in a day, and on Monday you start again from nothing. Pinterest pays you slowly and then keeps paying, because a pin is an entry in a search index rather than an item in a feed, and the ones you made last spring are still sending people to your site this morning while you were asleep.
Twelve questions, about three minutes, and no email address needed to see your score.
The shape is illustrative. The shape is also the entire reason this platform is worth your time.
What we run across Pinterest
The whole of it in one place, because the search research and the pin design are the same job done twice if you split them.
Where we come in
Different words each time, and underneath it is usually one of these.
Different words each time, and underneath it is usually one of these.
A designer made them, they look better than anything a competitor has posted, and the title says something like "New In" while the description is nine hashtags. Pinterest is a search index. If the words are not there, the picture may as well not be either, and this is by far the most common thing we are shown.
Three months of pinning, a graph that barely moved, and a decision made in the one window where the numbers always look disappointing. Pinterest pays late. Everything published in those three months was still ahead of them, and they turned it off two months before it would have started to show.
Pinterest counts an outbound click. Your analytics counts a session that survived the redirect. The gap is real, it is large, and nobody warned anybody about it, so the whole channel gets judged as a fraud when it is simply being counted twice in two different ways. Say it out loud in month one and it never becomes an argument.
The work is good, the post does what it does in about eleven hours, and then the counter resets and you owe the algorithm another one. That is the deal on every feed platform and it never improves with scale. Pinterest is the only mainstream place where this week's work is still earning next spring, which is why creators who crack it stop dreading Mondays.
The best post on your site has a single pin attached to it, made the week it went out, aimed at one phrase. There are four or five genuinely different searches behind that article and each of them wants a different picture and a different title. Most of the work is already done. What is missing is the four other doors into it.
The sessions are in the analytics, Pinterest is credited, and beyond that it is fog: which pin, which board, which search, which of them is worth making forty more of. Without that, every month is a guess dressed up as a strategy, and the pins that actually work get made once by accident and never repeated.
The Pin Lab
Nobody sees your pin on its own. They see it in a column, at about the size of a playing card, beside four other people who wanted the same search. Type into the boxes on the left and the grid on the right redraws with your pin in it, cut off in the places Pinterest cuts it. Change the shape and watch how much of the screen you win or throw away. The pin sitting in there now is a deliberately bad one, so you can watch every check fail.
0 of five checks clear. It is a checklist rather than a prediction: nothing can tell you in advance which pin the index will decide it likes.
All of this happens in your browser. What you type is not stored, not sent and not seen by us.
What the work actually is
Pinterest work is usually sold as design and it is mostly research and repetition. The picture matters enormously, but it matters second, after somebody has worked out which searches are worth entering and what those searchers are actually asking for.
What people type on Pinterest is not what they type into Google, and the guided search bar tells you for free. We build the phrase list before anybody opens a design tool, because it decides the board names, the titles, the descriptions and which pins get made at all.
Standing images at 2:3, text on the image that survives being shrunk to the width of a thumb, and several genuinely different pins per page rather than one pin resized four times. Templates so the volume is possible, variation so the volume is worth having.
Boards named after searches rather than after moods, sections that make sense to a stranger, the dead ones archived and the descriptions written as sentences. It is unglamorous, it takes a day, and it changes what the index thinks the account is about.
The feed connected properly so every product becomes a pin with a live price and stock level, rich pins turned on, and the whole catalogue eligible for shopping surfaces rather than sitting behind a button nobody presses. Mostly a plumbing job with a large payoff.
Pinterest Ads are cheap by the standards of every other auction and they are not a substitute for the organic work, they are a way of finding out in a fortnight what would otherwise take a quarter. We use them to test which pins deserve to be made forty more times.
The tag installed, conversions defined, and reporting built for a channel where the save and the purchase can be six weeks apart. Judge Pinterest on last click and you will switch it off. That is not Pinterest being weak, it is the report asking the wrong question.
If you only ever buy one of these, buy the first. Search research is the cheapest item on the list and it is the one that decides whether the other five were worth doing.
The Pinterest Scorecard
It is the same set of questions we work through on a first call, so you may as well have the answers before you decide whether to book one. Nothing is sent anywhere while you answer, and you see your score without handing over an email address.
Four things, weighted evenly, because a weak one drags the other three down with it. An account can be the most beautiful thing on the platform and fail on three of them.
Your answers stay in your browser. Nothing is stored, nothing is sent and your score never leaves this page, so copy it out below if you want to keep it.
The Vision Note
What is actually working in search, paid and AI right now, written for people who have a business to run rather than a marketing degree to finish. Sign up and the guide below comes through straight away.
Two of the twelve workflows in it are the ones this page keeps circling: turning one article into a dozen genuinely different pin titles and descriptions built around real searches, and turning a month of Pinterest data into one page somebody outside marketing will read. They work the same whether an agency runs your account or you do.
No more than twice a month, from [email protected]. Unsubscribe in one click, and your address is handled as set out in our privacy notice.
AI, used where it belongs
This platform rewards making a lot of things and it punishes making the same thing repeatedly. For most of its history that was an impossible pair of instructions for anybody without a design team. It is not any more, provided the machine is used for the drafting and the repetition rather than for the judgement, and provided a person still decides what is worth publishing. The moment nobody is deciding, you have built a very efficient way of filling an index with content nobody wanted.
How we use AI across the agencyA single article usually answers four or five different searches and most sites have one pin pointing at it. Drafting a dozen titles and descriptions, each aimed at a different phrase, is an hour instead of a week. A person cuts it to the four that are honestly different and the rest go in the bin.
Descriptions rewritten out of the words a company uses about itself and into the words a person types when they are planning a kitchen. It is the same job as rewriting a product feed and it has the same effect, which is that the index finally understands what the thing is.
Four real alternatives of a pin, with a different image, a different crop and a different promise, so the test means something. Four exports of the same layout with the colour changed is not a test, it is the same pin four times, and the index treats it accordingly.
We will not put an invented room, an invented product or an invented person on a pin that leads to a real shop. It is the fastest way to lose somebody at the landing page, it is dishonest, and on a platform where people are planning something they intend to buy it is worse than useless.
What we measure
Not whether it works. When it starts, when it stops, and what a month of it is still worth a year later. Put your own numbers into the panel below and it will tell you the honest version, including the case where you should not start at all.
A pin does not expire, but it does fade, and the fade is slow enough that this month's traffic is mostly the work of previous months. Move the sliders to something close to your own situation. The number on the right of the panel is the one that matters: how much of a given month's traffic you had already paid for before that month began.
Move a slider and this panel says whether the arithmetic holds.
Across the whole run that is 30,126 visits, worth roughly £4,519 at your own figure.
Stop completely at the end of it and a year later there are still about 1,615 arriving every month, from work you had already finished and forgotten.
Straight arithmetic rather than a forecast. It assumes every pin behaves like the average one, which none of them do: in practice a handful carry most of the traffic and the rest do very little, which is the argument for volume rather than against it. The point of the panel is the second number and the month you stop, not the decimals.
If the report cannot tell you which individual pins are earning, you are not running Pinterest. You are decorating it.
An illustration of the monthly page rather than a real client's numbers. The order is the argument: the second row is the one nobody else reports and the one that tells you whether any of this is compounding.
How it runs
The first four weeks are mostly research and plumbing, which is unsatisfying to buy and is the reason most Pinterest accounts never work.
What people actually type on Pinterest around what you do, pulled from the guided search, the trends tool and whatever your own analytics already knows. You get a list of phrases, grouped, with a note beside each one saying whether it is worth entering and what would have to be true to win it.
Site claimed, rich pins working, the tag installed and tested, boards renamed and rewritten around the phrase list, the dead ones archived. Then the back catalogue: the pages you already have that deserve pins and never got them, which is usually where the first three months of traffic comes from.
New pins every week, in batches, aimed at named phrases and pointed at pages that can take the traffic. Several genuinely different pins per page rather than one resized. The ones that move get another four made like them, and the ones that do not are quietly dropped.
One page, twenty minutes, in plain English. Sessions rather than impressions, which pins earned them, how much of the month came from earlier work, and what we are making more of as a result. Month one and month two will look thin. We will say so before you notice.
Monthly is the common one, and for most businesses it should be bought in blocks of six months rather than three, because three is exactly long enough to pay for the setup and see none of the return.
Monthly is the common one. A fair number of creators are better off buying a day, taking the phrase list and the templates, and doing the publishing themselves, and we would rather say that than sell you a retainer.
The research, the foundations and the templates, done with you, for anybody who intends to keep publishing themselves.
We research it, make the pins and publish them, with one senior person named on the account.
A defined piece of work with a start and a finish, quoted on one page before it begins.
Everything is quoted before anything starts, so you see the scope and the number on one page and decide from there. Any advertising budget is separate and stays on your own card, in your own account, where you can see it. We also work quietly behind consultants and smaller agencies who want the Pinterest half handled properly without losing the client relationship, so if that is you, say so on the call.
Before you ask
It behaves like a search engine and it is sold like a social network, which is why so many accounts are run wrongly. People arrive with something in mind and type it in. They are not there to see what their friends did at the weekend, there is no meaningful concept of a feed you have to keep feeding, and a pin from 2023 will happily outrank a pin from this morning if it answers the search better. Once you accept that, the work stops being about posting and starts being about which searches you intend to be present for.
Something usually moves inside six to eight weeks and none of it is worth judging before month four. The reason is structural rather than an excuse: a pin builds slowly, so almost everything you publish in the first quarter delivers most of its traffic in the second and third. That is also why the most expensive mistake on Pinterest is stopping at month three, having paid for all of the setup and collected none of the return. If six months is not available, we would rather sell you a single day and let you run it.
Those categories are enormous there, which is where the reputation comes from, but the useful test is not the category. It is whether anybody plans the thing before they buy it. Kitchens, holidays, garden rooms, financial milestones, home education, business software for people who are about to start a company, anything with a project attached to it. If your customers research for weeks before they commit, there is a version of this that works. If they buy on impulse in four minutes, there probably is not, and we will say so.
Because they are counting different events. Pinterest counts an outbound click leaving the platform, your analytics counts a session arriving and staying long enough to register, and a meaningful share of clicks never complete that journey, especially inside the in-app browser on a phone. The gap can be forty per cent or more and it is not fraud. Agree at the start which number you are managing to, put both on the report, and it never becomes an argument in month five.
No, and they have not for years, but they are still all over most accounts because a well-meaning article in 2019 said they did. The words that do the work are the ones in the pin title, the description written as an actual sentence, the board name, the board description and the text on the image itself. Replacing nine hashtags with one honest sentence describing what the picture is, in the words somebody would have typed, is the single most common improvement we make.
More than feels reasonable and fewer than the automation tools will sell you. In practice, somewhere between fifteen and forty a month keeps an account genuinely present without turning into a duplicate factory, and the shape of it matters more than the count: several different pins for the same page, aimed at different searches, spread over weeks rather than dumped in a batch. Volume made of near-identical pins is worse than half the volume made of different ones.
That was standard advice for years and it stopped working. Pinterest now heavily prefers what it calls fresh pins, which means a new image pointing at a URL rather than the same pin recirculated. It does not have to be a new photograph. A different crop, a different text treatment, a different angle on the same page is enough. But pinning the identical image to nine boards is a strategy from a version of the platform that no longer exists.
Sometimes, and rarely as the main event. Clicks are cheap compared with almost any other auction, which makes ads unusually good for finding out quickly which pins and which phrases deserve more work, rather than for buying traffic at scale. The failure we see is a company running ads against an account with no organic foundation underneath it, which is like buying billboards for a shop with no sign on the door. Fix the free part first, then use the paid part to accelerate what is already working.
They stay yours and they keep working, which is a genuinely unusual thing to be able to say. The account is in your name, the boards are yours, the pin files and templates are handed over, and the traffic does not switch off when the invoices do. It fades over a year or two rather than stopping on the last day of the month. That is a straightforward argument for the channel and it is also, honestly, an argument for not needing us forever.
No. The scorecard, the pin lab and the calculator all run entirely in your browser, your answers are gone the moment you close the tab, and your score is never sent to us at all. There is a button under your result that copies it out as text if you want to keep it. If you sign up for the guide further down the page, that is a separate thing and only your email address travels, to Brevo, which sends our email. It is all set out in the privacy notice.
Also worth a look
The same treatment for the other platforms, each with the tool that only makes sense on that one.
Book a call
Share your Pinterest account and your analytics on the screen and we will go through them with you. You will leave with at least one thing worth changing this week, whether or not you ever hire us.
Book a call