You spent an hour tearing apart a pricing page, and the founder skimmed it for ten seconds and moved on. The problem is almost never the findings. It is how they were packaged. A pricing teardown gets ignored for a short list of avoidable reasons, and each one has a fix you can apply before you hit Publish.
The job of a pricing page teardown is to let a buyer, or the founder reading your review, tell what they would pay and which plan is theirs. If your review buries that, the rest does not matter.
You dropped pins but never said what they point at
A screenshot with three numbered pins and no matching notes is a puzzle. The founder sees a dot on the pricing table and has to guess whether you mean the price, the feature row, or the toggle next to it. It is tempting because dropping a pin feels like the work is done. It is not.
Fix it by making every pin resolve to a plain comment. Pin 1 points at the annual toggle; the note says what is wrong with it. In Cobalt Capture the pins and drawings are baked into the exported image, so whoever opens the review sees the same marked screen you did. There is good detail on pinning the exact spot you mean and on ordering numbered pins so the reader walks the page in sequence.
You cropped so tight the page lost its context
A cropped shot of a single $49 price tag, with nothing around it, tells the reader almost nothing. Is that per user? Per month? The middle plan or the top one? Cropping is tempting because it looks focused, but a price only means something next to the plan name, the billing period, and the plans on either side of it.
Crop to the decision, not to the number. For a pricing teardown that usually means the whole plan card or the full comparison row, including the headers that label the columns. If you are unsure, the piece on why a cropped still beats an annotated live page covers where to draw the line. Leave in enough that the founder recognizes their own page at a glance.
Your comments run three paragraphs each
A busy founder will not read an essay under every screenshot. Long comments feel thorough while you write them, and they are the first thing skipped. The reader wants the finding and the cost of it, not your reasoning in full.
Write each comment as the problem and what it costs the buyer. "A solo user cannot tell if the $49 plan is per seat. They will assume the worst and leave." That is one line and it lands. If you have a lot to say and would rather talk it through, dictate the comment instead of typing; Cobalt uses the browser's built-in speech recognition in Chrome and Edge, and there is a case for dictating a long note and then trimming it.
You never worked out what a real buyer would pay
"The pricing is confusing" is an opinion. "A team of 5 lands on $245 a month but the page only shows the per-seat number, so they do the math wrong and bounce" is a finding. Skipping the math is tempting because it is tedious, but the whole point of a pricing review is whether a buyer can tell what they will actually pay.
Run the cost for a solo user, a team of 5, and a company of 25, and put each number in the review. Check the fine print and the phone layout too, because the plan that reads clearly on desktop often collapses into an unreadable stack on a 375px screen. If you would rather have this done for you, you can ask your AI assistant to run the study end to end; the ready-made pricing teardown works out those three costs and checks the phone view for you.
You led with the small stuff and buried the verdict
If your review opens with a note about button color and the actual problem (nobody can tell which plan is theirs) is halfway down, the founder has already closed the tab. Ordering findings by what you noticed first is the default, and it is wrong.
Lead with the verdict and rank everything below it by what it costs. The clearest signal that a plan is wrong goes first; the cosmetic note goes last, or gets cut. A teardown that opens with the one thing to fix gets acted on. The format that a founder reads on their phone is the same shape: verdict up top, costs stated, specifics underneath.
You made the founder sign up to read it
A review locked behind a login is a review nobody opens. When you Publish in Cobalt, the review gets a short link anyone can read, plus a clean reader page for people who just want the findings. No install, no extension, no password for the person opening it. You can try the whole thing free with a new review, and unlimited reviews and screenshots cost nothing until you need the editing and export to stay open past the seven-day window; after that, Pro is $5 a month and reopens any read-only review instantly.
Pick one of these you keep doing and fix it on your next teardown. Rank by cost, crop to the decision, and keep every comment to the finding and the price it carries. The founder reads the whole thing and knows exactly what to change.