Product & Mechanism ad angles

The ad that stops arguing and shows. No mechanism, no explanation, no adjectives — the product doing the thing, in a frame the reader can check. It is the only angle in the library where the evidence and the explanation are the same shot, which is why the taxonomy lists it as strong at every stage and why it is the most under-tested ad most accounts have.

What you show matters far more than how well you shoot it. Show the moment of change: the input going in, the result coming out, ideally without a cut. Continuity is the whole proof — every edit is a place where the reader assumes something was fixed off-camera, and a demo assembled from seven flattering fragments persuades less than one unglamorous continuous take. Resist the interface tour, too. A tour is a list of features arranged as a video, and it answers a question nobody asked.

Its known weakness is real: if the result isn’t visually interesting, a demo is boring. The fix is usually to change what you’re demonstrating rather than to abandon the angle. When the outcome is invisible — software, services, anything financial — demo the input instead, because how little the customer has to do is itself a visible result. Two fields and a button is a demonstration. So is the speed at which a screen fills in.

Input, mechanism, result — the path the customer will walk, laid out end to end. Where a demo shows one instance, this describes the shape of every instance, which is what someone needs before they can picture themselves doing it. Its job is uncertainty reduction, and uncertainty is the quiet reason most people who liked your benefit ad still didn’t buy.

The reader is not trying to learn your system. They are trying to price the effort, and every step you name is a step they mentally add to their week. So name three. Three is enough to feel like a real process and few enough to feel survivable; five reads as a project, and a project gets postponed. The other rule follows from the same place: every step must be something the customer does or something that visibly happens to them. Internal stages of your pipeline are not steps, they are your problem, and putting them in the ad makes your product sound like work.

It becomes too technical exactly when the writer starts enjoying it. The test is to read the ad as someone deciding whether to spend a Saturday on this. If a step needs a term the reader doesn’t already own, either teach the term in four words or cut the step — and if the honest process really is nine steps, the ad’s job is to tell them which two they’re responsible for.

Not what it does — why it can do it when the obvious alternative can’t. This is the reason-to-believe angle proper, and the one that does double duty: a mechanism is simultaneously an explanation and a differentiator, which is why it is worth more test budget than its position in the library suggests.

It requires a difference you can name, and the tell that you have a real one is that it also explains a limitation. A mechanism that produces nothing but upsides is a slogan wearing a lab coat. “It reads your website instead of asking you to fill in a brief” is a mechanism — and it explains why the setup is quicker and why a business with a thin website gets a thinner first draft. Saying the second half is not a concession; it is the thing that makes the first half believable, because a reader who has been sold to before knows every real design decision cost somebody something.

Sequence matters more here than anywhere else in the family. Name the mechanism in the first line and it is jargon; name it after the reader wants the outcome and it is a brand. So lead with the consequence they can feel, reveal the machinery second, and only then attach a capitalised name to it if you intend to keep using that name for years. And be honest with yourself about whether you have one: if the mechanism ad underperforms a plain benefit ad at the same audience, the market has just told you your difference is invisible from outside, which is a positioning problem no headline will solve.

One specification, immediately translated. The feature supplies credibility because it is concrete and checkable; the benefit supplies motivation because it happens in the reader’s life rather than in your product. Neither works alone — a feature by itself is trivia, a benefit by itself is the unbacked promise issue #2 spent ten angles trying to rescue — and the translation between them is the entire ad.

Write it as one sentence with a hinge in the middle. Four sizes in one ad, which means Stories stops cropping your product out of frame. Every ad lands paused, which means you can generate at midnight and decide in the morning. The hinge is where most of these fail: the writer states the feature, feels the relief of having been specific, and stops. If the ad ends before “which means”, you have published a spec sheet with a logo on it.

The taxonomy’s warning — that it still becomes a catalogue if overused — is really a warning about quantity. One feature per ad. The moment there are three, the reader stops translating and starts skimming, and you have written a Benefits List with worse pacing. There’s a strategic side effect worth using deliberately: the benefit you attach declares who the ad is for. The same export feature translated as “hand it to your client by Friday” and as “stop paying for a designer to resize things” is two different ads aimed at two different buyers, and running both is a cheap way to find out which one you actually have.

One input carries the whole argument. The fabric, the compound, the chip, the source of the data. It is the premium-positioning workhorse in physical categories because an ingredient is a quality claim you can point at, and pointing is worth more than describing.

The taxonomy’s weakness — the ingredient may mean nothing without education — is not a caveat, it is the job. Budget two sentences: what it is, and what it does that the ordinary alternative doesn’t. A quantity beats an adjective every time, because a number implies someone chose it and a choice implies a trade-off. “Merino” is a word the reader may or may not have opinions about; “230gsm merino, so it doesn’t hold smell for a week of wear” is an argument with a reason attached.

There is a second effect worth naming, because it is why brands pay for ingredients they could have made themselves: a known input transfers credibility you didn’t have to build. That is also its risk. Lean on someone else’s ingredient hard enough and you have taught your market to shop for the ingredient, which is available to your competitors at the same price it was available to you. The safest version is an input that is genuinely yours, or a combination that is — and the fastest way to find out whether the education landed is whether customers start repeating the ingredient back to you unprompted.

The advantage is that it’s new. This runs colder than most of the family because novelty needs no prior belief — new is interesting on arrival, and interest is the thing a cold ad is short of. It is also the angle that has degraded fastest in the last three years, for a reason worth being blunt about.

“AI-powered” is not a differentiator any more; it is a table stake, and an ad whose entire technological claim is that it uses AI is an ad about the year rather than about the product. The same was true of “cloud” and “app” before it. The fix is ordering: lead with the capability that didn’t exist before, name the technology second, and only if naming it adds something the capability didn’t already prove. Jargon is not a vocabulary problem, it is a sequencing problem — the identical term reads as substance after a consequence and as noise before one.

Novelty also has a shelf life, and the ad expires when the category catches up. So treat a technology ad as a claim with a clock on it: run it while the difference is real, watch for the month the mechanism becomes ordinary, and have the durable version — a mechanism, a benefit, a proof — already tested and waiting. Anything you plan to still be selling in two years should not depend on a word that will sound dated by then.

How it was made, offered as proof of what it is. The reader can’t evaluate the product from an ad — they have never held it, never used it — so they evaluate the process instead and generalise from that. Care visible in the making is taken as evidence of care in the result, and that inference is what lets a craftsmanship ad support a price a spec sheet couldn’t.

Specific and unglamorous beats beautiful. The slow-motion beauty shot is what everyone shoots; the detail that actually moves people is the one with a cost in it — the batch that gets rejected, the fourth pass nobody would notice was missing, the supplier you kept when a cheaper one appeared. Anyone can claim to care. Only someone who does can name the thing that caring cost them, which is why a trade-off is evidence and a sentence about passion is decoration.

The taxonomy is right that this is production-intensive and weak for commodities, and it looks at first like it doesn’t apply to software at all. It does — it just isn’t “our engineers are passionate”. The software version of craftsmanship is the decisions you made against your own short-term interest: the default that protects the user, the feature you refused to ship, the thing that is slower because the fast version was worse. Those are checkable, they are the reasons your product feels different in use, and they are almost never in anyone’s advertising.

Not “does it work” but “will it keep working”. That question only arrives once someone is close enough to imagine owning the thing, which is why this sits Mid–Hot and why it converts: it is often the last live objection between a decided buyer and a payment.

It has to be demonstrated, not asserted. Every category has its version of the torture test — the bag dragged behind a car, the ten-year warranty, the machine on its four hundredth cycle. For anything digital the equivalent is consistency: the hundredth run looking like the first is the claim, because software rarely breaks, it drifts. What makes any of these persuasive is that they are falsifiable. A number with a unit and a duration attached invites the reader to check it, and inviting the check is most of the persuasion.

Then live with it, because this is the only angle in the family your customer re-audits daily. A durability claim doesn’t end at purchase; it becomes the standard your product is measured against every time it is used, and the gap between the ad and the third bad week is where refunds and one-star reviews come from. Which is the useful discipline: only make the reliability claim you would still be comfortable with printed on the box a year in.

It does more than you assumed. The argument is arithmetic on perceived value — one price, several jobs — and it is at its best when the second use is one the reader was already paying somebody else for.

Its weakness is the most predictable in the family and the most ignored: say everything and you stand for nothing. The reader files products under one heading, and an ad offering eight uses gets filed under none. So take the structure seriously — one primary use, one surprising second, and stop. The second should be adjacent enough to be believed and far enough to be news, which is a narrower target than it sounds. A blender that also crushes ice is not news. A blender that grinds coffee is.

There is a better place for this angle than cold traffic, and it’s the one most accounts skip: existing customers. Someone who already owns the product and doesn’t know about its second use is the cheapest expansion audience you have, and the ad is doing a retention job while it runs. Cold, be more careful — a stranger who hasn’t agreed you’re good at the first thing has no reason to be impressed that you also do a second.

One capability, held back and then revealed. It borrows the mechanics of curiosity — an open loop, a small withholding — but unlike a pure curiosity hook it is paid off by the product itself, which is what keeps it honest and what makes it convert instead of merely getting clicked.

The reveal has to happen inside the ad. A gap you only close after signup is clickbait with a product attached, and the bounce arrives about eleven seconds later carrying your acquisition cost with it. So pick something demonstrable in one frame or one sentence, and make sure the surprise is genuinely a surprise: the best source isn’t the roadmap, it’s support tickets that begin “wait, it does that?”. Those sentences are your market telling you which capability is invisible from outside, and invisible-but-loved is exactly the raw material this angle runs on.

Novelty wears off, as the taxonomy says, so treat this as a rotation rather than a position — one reveal per quarter, retired the moment it stops feeling like news. One caution about what you choose: an aha that turns out to be a niche capability produces a very engaged audience that doesn’t buy. The feature should be unexpected in the same product the reader was already considering, not a reason to reconsider what the product is.