There is a moment in every services pitch where you ask the merchant to believe something about you that they have no evidence for. Most agencies try to bridge that gap with proof borrowed from other people — case studies, logos, testimonials. It works, slowly, and only if the merchant decides those other people resemble them.
The faster bridge is to do a small piece of the work before anybody has paid you, and simply hand it over. Not a proposal. Not a deck. A teardown: a short, specific, slightly uncomfortable document about their store, built from real data, that would be useful to them even if they never speak to you again.
What a teardown is, and what it is not
A teardown is not an audit. An audit is a hundred-page instrument of torture that costs you three days, tells the merchant forty things, and gets skimmed exactly once. A teardown is one to three pages, takes you under thirty minutes at the margin, tells the merchant three things, and gets read to the end.
The distinction that matters is the reader. An audit is written to demonstrate thoroughness — it is a document about you. A teardown is written to be acted on this week — it is a document about them. If a merchant could forward your teardown to a developer and get something done from it, you built it correctly.
It also has to be honest, including about what is working. Opening with two or three things the store does well is not politeness, it is credibility management. A document claiming everything is broken reads as a sales tool, and merchants have seen enough of those to discount them on sight.
The five blocks
Use the same structure every time. Consistency is what makes this scalable, and it is why the tenth teardown is sharper than the first.
Block one, the snapshot: what the store is, in their language, from public data. Platform, catalogue size, countries served, current organic traffic and keyword coverage, Google Business rating and review count if a profile exists. This block proves you looked, and it takes zero minutes of your time because it is a data pull.
Block two, the trajectory: where the numbers have been going. This is the block nobody else has, because it requires history. With roughly 170 monthly buckets of organic traffic and position data per store, you can write “your organic sessions peaked eighteen months ago and have declined in eleven of the last fourteen months” — a sentence that stops a merchant cold, because they have felt it without ever seeing it stated.
Block three, the diagnosis: one plausible cause, not five. You are not certain and you should say so. “The decline is concentrated in category pages rather than product pages, which usually means…” is far more persuasive than a confident list, because it invites the merchant to correct you — and a merchant who corrects you is a merchant in conversation.
Block four, the three moves: prioritised, with effort and expected impact stated in plain language. Not thirty recommendations. Three, ranked. One of them should be something they can do themselves this week without you, which feels counterintuitive and is the single highest-trust move in the entire document.
Block five, the boundary: what you did not check, and what access you would need in order to check it. Analytics, search console, the real conversion rate, their margin structure. This is where the engagement lives — you have demonstrated competence on public data and named precisely what the private data would unlock.
Where the content comes from
The reason teardowns do not scale for most people is that they are assembled by hand from six browser tabs. Assembled from a single API call, they scale perfectly well.
The fields that carry a teardown are unglamorous: domain, platform, country, merchant name, product count, monthly traffic, organic keyword count, AI mention and citation counts, currency and language, the detected app and tech stack, payment methods, shipping carriers, socials, and — where a Google Business profile matched — category, rating, review count, photo count and whether the listing is claimed. Add the monthly trend series and every sentence of blocks one and two is written for you.
The judgement lives in blocks three and four, and that is exactly right. You are not trying to automate the thinking. You are trying to remove the forty minutes of data gathering that currently stops you doing the thinking thirty times a week instead of three.
Making it repeatable
Write the template once, as a document with slots. Then decide, per segment, which slots matter. A teardown for a declining-traffic Shopify store in fashion needs different blocks three and four than one for a WooCommerce store with healthy traffic and a 3.8-star Google profile. Two or three template variants cover most of a serious ICP.
Generation is the easy part. Run an agent against the store record with the template and a strict instruction to cite the field behind every claim, and you get a solid draft. The rule that keeps it usable: every sentence in blocks one and two must trace back to a field, and the agent is forbidden from asserting anything it cannot cite. Then you spend ten minutes on blocks three and four, which is the only part that ever needed you.
Version the template. When a teardown converts, note which sentence the merchant quoted back at you in their reply — merchants almost always quote one line. That line is your hook, and after twenty teardowns you will have three or four lines that reliably provoke a response.
Delivering it
Attach it. Do not put it behind a form, a calendar link, or a “happy to share it if you’re interested”. The entire mechanism depends on the merchant receiving something of value before deciding anything, and every gate you add converts a gift back into a trade.
Keep the covering message shorter than feels comfortable. Three or four lines: the single most striking finding, the fact that the document is attached, and one question. The question should be answerable in a sentence, and it should not be “would you like to book a call”. “Does the decline line up with anything you changed last spring?” gets answered. “Do you have 15 minutes Thursday?” gets ignored.
A short screen recording walking through the document lifts response meaningfully, at the cost of a few minutes per prospect. Reserve it for your top segment. For everybody else, the document does the work on its own.
Turning a teardown into an engagement
The reply you get is rarely “yes, let’s work together”. It is usually a correction, a question, or a complaint about their developer. All three are the same thing: an opening. The teardown has moved you from vendor to person-who-knows-something-about-my-store, and the next step is not a pitch. It is a conversation about block five.
The natural sale from a teardown is the paid diagnostic — a fixed-price, short engagement where you get the access you named and produce the real plan. It is easy to say yes to, it filters out non-buyers, it gets money moving, and it turns the retainer conversation into a continuation rather than a cold start. Merchants who have already paid you once evaluate the second proposal in a completely different frame.
Do not skip that step in order to sell the retainer directly. The teardown earned you attention, not confidence in your ability to execute for six months. Ask for the thing the document actually justifies, and let the document justify the next thing later.
The ways this goes wrong
It goes wrong when the teardown is a brochure with a data section stapled to it. Merchants can smell a sales document, and the moment one finding is visibly bent to make your service the obvious answer, everything before it becomes suspect. Report what the data says, including the times it says the store is fine.
It goes wrong when the volume outruns the judgement. Fifty auto-generated teardowns with an identical block four are fifty pieces of evidence that you do not think. Twelve with a genuine opinion in them will out-earn them every single time, and they take less of your week.
And it goes wrong when it stops. The teardown is not a campaign, it is a standing artifact — the thing you send instead of a pitch, permanently, to every prospect, in every channel, for years. Its entire advantage is that almost nobody else is willing to do the work before the invoice. That advantage compounds, quietly, for as long as you keep sending them.