You can run a useful demand-generation teardown without account access by auditing the public path from ad to promise, offer, proof, page experience, qualification, and next step. You cannot honestly diagnose spend efficiency, bidding, targeting, attribution, lead quality, or revenue performance from that evidence alone.
That distinction makes the teardown more valuable, not less. A public review can find message discontinuity, vague offers, unsupported claims, hidden fit requirements, dead-end calls to action, and unnecessary friction. It can produce a sharp test brief. It should not manufacture a media-buying story from a few visible ads.
At Stackmatix, the growth agency I co-founded, I want acquisition analysis to connect with qualified customers and sales outcomes. This template reflects that standard. The worked company, observations, and scores below are hypothetical; they are not client results or benchmarks. Public platform documentation was checked on September 15, 2026.
What can a public demand-generation teardown establish?
Start by labeling every statement as observed, inferred, or unknown.
| Evidence class | What belongs here | How to write it |
|---|---|---|
| Observed | Visible ad, public page, form field, price, claim, proof item, navigation, or technical behavior | “The landing page asks for eight required fields.” |
| Inferred | A plausible strategic explanation for several observations | “The company may be prioritizing enterprise evaluation.” |
| Unknown | Spend, audience settings, bidding, conversion rate, lead quality, sales follow-up, pipeline, or retention | “Public evidence does not show whether the campaign is efficient.” |
Do not turn “I saw three ads” into “this is the company's strategy.” The ads may represent one geography, format, date range, experiment, or subset of the inventory. A public teardown is a review of the visible buying experience, not an audit of the advertising account.
Google describes its Ads Transparency Center as a searchable repository where people can look up advertisers and ads served across Google platforms and filter by factors such as date and targeted location. That makes it a useful observation source, not a performance report. Google Ads transparency documentation and Google Ads Transparency Center.
Where available, also inspect the official Meta Ad Library and LinkedIn Ads Library. Record the country, date, advertiser identity, filters, and URLs you used. Platform availability and visible fields can change.
Define the audience and buying job before scoring
A teardown without a reader definition becomes a design critique. Write one sentence first:
A [specific person] in [specific situation] needs to decide whether [offer] can help them [job] under [important constraint].
If the company appears to serve several audiences, choose one journey. Do not average a self-serve developer path, an enterprise security evaluation, and an executive strategy offer into one score.
Record the intended action too. A trial, free tool, technical evaluation, webinar, demo request, and direct purchase create different burdens of proof. The page should help the reader make the decision actually being requested.
The ideal-customer-profile worksheet helps translate a broad audience label into a workflow, trigger, constraint, and buying condition.
Use the PUBLIC teardown framework
The framework covers six observable layers:
| Layer | Question | Evidence to collect |
|---|---|---|
| P — Public inventory | What messages and destinations are visible now? | Ad-library records, public search results, social posts, pages, forms, and calls to action |
| U — User and job | Can the intended buyer recognize their situation? | Role, trigger, problem, workflow, environment, and exclusions |
| B — Buying offer | Is the proposed exchange concrete? | Outcome, product or service, scope, effort, price or commercial path, and next commitment |
| L — Linked promise path | Does each step continue the same argument? | Ad claim, headline, body, proof, CTA, form, confirmation, and follow-up expectation |
| I — Integrity and proof | Does the evidence support the nearby claim? | Product demonstration, sourced metric, customer statement, case study, documentation, and qualifications |
| C — Conversion and qualification | Can a suitable buyer act and a poor fit self-select? | Form purpose, required fields, eligibility, mobile experience, success state, and next owner |
Score each layer from 0 to 2:
- 0 — Missing or contradictory: the layer prevents a clear decision.
- 1 — Partial: useful elements exist, but an important question remains unresolved.
- 2 — Coherent: the public evidence answers the relevant question for this journey.
The total is a prioritization device, not a universal benchmark. Save the observation behind every score so another reviewer can disagree productively.
1. Capture the public inventory without cherry-picking
Create a dated evidence log before interpreting anything.
| Field | Entry |
|---|---|
| Review date and timezone | When the observation was made |
| Advertiser or company | Exact public identity |
| Market and device | Country, language, desktop or mobile |
| Source | Official ad library, public search, website, social page, or product documentation |
| Ad or page URL | Stable link where available |
| Message | Exact claim or a concise faithful summary |
| Creative format | Text, image, video, document, or other visible format |
| Destination | Final public page reached |
| Intended action | Trial, signup, demo, download, purchase, or another step |
| Status | Observed, inferred, or unknown |
| Screenshot or note | Evidence captured within applicable platform rules |
Collect meaningful variation rather than the first three items. Look for different problems, audiences, offers, proof types, formats, and destinations. If the inventory looks repetitive, record that as an observation rather than assuming the account lacks other tests.
2. Test whether the user and job are recognizable
Read the ad and first screen of the destination as a buyer would. Can the person identify:
- Who the offer is for?
- Which problem or trigger makes it relevant now?
- What changes in the workflow?
- Which product, environment, company stage, or use case is supported?
- Which obvious situations are not a fit?
“AI-powered growth for modern teams” names neither the work nor the buyer. A narrower line such as “Review pull requests against your team's coding standards” gives an engineering audience a task to evaluate. Specificity can reduce raw response while improving the usefulness of the response.
Do not confuse jargon with qualification. A phrase can sound technical and still leave the use case undefined.
3. Write the buying offer in one row
Reduce the offer to six fields:
| Field | Question |
|---|---|
| Customer | Who should act? |
| Problem | What costly or important condition exists? |
| Outcome | What becomes better or possible? |
| Mechanism | What product or workflow produces that change? |
| Commitment | What must the buyer provide now—time, data, access, money, or internal coordination? |
| Next step | What decision does the CTA initiate? |
If you cannot complete the row from public information, the buyer may face the same problem. The fix is not automatically more copy. It may be a tighter offer, a more appropriate CTA, or visible implementation requirements.
Use the startup offer playbook when the exchange itself remains unclear.
4. Trace the linked promise path
Make one row for each public journey:
| Step | Promise or expectation | Required action | Friction or contradiction |
|---|---|---|---|
| Ad | Why the person should care | Click or engage | Does the message imply unsupported scope? |
| Landing-page opening | What the offer is | Continue reading or start | Does it repeat or abandon the ad's argument? |
| Proof | Why the claim is credible | Evaluate evidence | Does the proof support this exact claim and audience? |
| CTA | What happens next | Trial, form, purchase, or contact | Is the commitment proportionate to the evidence? |
| Form or product start | What information or setup is needed | Submit or begin | Are purpose, privacy, eligibility, and effort clear? |
| Confirmation | What the company will do | Wait, schedule, verify, or use | Is there a clear success state and owner? |
The handoff matters. An ad promoting a template that lands on a generic demo page breaks the promise even if both assets look polished. A page advertising immediate access that ends with “we'll be in touch” creates a different mismatch.
The search-ad copy checklist shows how to review claims and destinations before launch.
5. Match proof to the claim
Inventory each material claim and the nearest supporting evidence.
| Claim type | Stronger public evidence | Common weak substitute |
|---|---|---|
| Product capability | Current product, documentation, or precise demonstration | Generic category language |
| Customer relationship | Approved logo, statement, or case study with clear scope | Unexplained logo wall |
| Performance result | Dated method, population, period, and attributable result | Percentage with no denominator or source |
| Comparative claim | Consistent criteria and current primary observations | Selective feature list |
| Security or compliance | Current official documentation and exact qualification | Vague trust language |
| Business outcome | Customer-approved evidence tied to the represented workflow | Product activity presented as revenue impact |
Treat a testimonial as evidence of what the named person said, not automatic proof of every claim around it. Treat a customer logo as evidence only when the relationship and permission are accurate.
The marketing-claim sourcing checklist provides a deeper evidence review.
6. Inspect conversion and qualification together
Reducing friction is useful when the removed work does not protect customer fit or execution. A two-field form can generate more submissions while giving sales too little context. A twelve-field form can repel a suitable buyer before the company has earned the information.
Ask:
- Does each required field have an immediate purpose?
- Are price, eligibility, platform, geography, company size, or implementation conditions visible when material?
- Does the CTA accurately name the next step?
- Can the buyer complete it on a phone without horizontal scrolling or broken controls?
- Does the success state explain timing and ownership?
- Can an unsuitable buyer recognize a better path without submitting?
Google's page-experience guidance recommends checking mobile display, secure delivery, intrusive interstitials, and whether the main content is easy to distinguish. Those checks are useful for the public experience; they do not predict the page's conversion rate. Google Search Central: page experience.
Use the paid-traffic landing-page review for a focused prelaunch checklist.
Work through a hypothetical teardown
Consider a fictional B2B SaaS company called Northline. It sells an AI workflow product to revenue-operations teams. The following ads, pages, and scores are entirely hypothetical.
The visible inventory contains three messages:
- “Turn call notes into an approved CRM action plan.”
- “Automate revenue operations with AI.”
- “See how fast-growing teams eliminate manual work.”
All three ads lead to one page headed “The AI Operating System for Revenue.” The page shows six customer logos, describes many possible workflows, and asks visitors to “Get Started.” That CTA opens an eight-field demo form. Pricing, supported CRM systems, required access, review responsibilities, and what happens after submission are not stated.
Here is the hypothetical score:
| PUBLIC layer | Score | Evidence |
|---|---|---|
| Public inventory | 2 | Three materially different messages and one shared destination are documented |
| User and job | 1 | Revenue operations is implied, but only the first ad names a concrete task |
| Buying offer | 0 | Outcome, implementation scope, and immediate commitment remain unclear |
| Linked promise path | 0 | The specific call-note workflow disappears on the general page |
| Integrity and proof | 1 | Logos are visible, but no proof is attached to the workflow claim |
| Conversion and qualification | 1 | The form works, but “Get Started” does not disclose a demo or key fit requirements |
| Total | 5 of 12 | Prioritize offer and promise continuity before cosmetic page changes |
The score does not show that Northline's campaigns perform poorly. It identifies a public decision problem: the strongest specific message is not continued or substantiated after the click.
Turn the teardown into one test brief
Choose the largest observable break with a plausible commercial consequence. For the hypothetical Northline journey:
- Observation: The call-note ad names a specific workflow; the destination does not.
- Unknown: Public evidence does not show spend, targeting, conversion rate, lead quality, or sales outcomes.
- Hypothesis: Continuing the workflow, buyer, supported environment, and next step will help suitable visitors evaluate fit.
- Change: Route that message to a page explaining the call-note-to-CRM workflow, required access, human approval, proof, and demo agenda.
- Primary measure: Qualified demo progression under the company's existing definition.
- Diagnostic measures: Page engagement, form start, completion, sales acceptance, and stated disqualification reasons.
- Guardrails: Qualified volume, cost per accepted opportunity, sales burden, and downstream activation.
- Decision rule: Continue, revise, or stop after the required buying-cycle window and sample are defined.
Those measurement fields require account access. The public teardown produces the test; the internal team determines whether the test works.
Use this final teardown template
Complete the review in this order:
- Define one buyer, situation, job, constraint, and intended action.
- Log visible ads and pages with dates, markets, devices, sources, and filters.
- Label every conclusion observed, inferred, or unknown.
- Score the six PUBLIC layers from 0 to 2 and attach evidence.
- Identify the first break in the buyer's decision path.
- Write one alternative offer or path that resolves that break.
- Specify which private account and CRM evidence is needed to validate the diagnosis.
- Create one test brief with a business outcome, guardrails, owner, and decision window.
A credible public teardown should make the next investigation or experiment sharper. It should never imply access you do not have. The standard is simple: show what is visible, say what remains unknown, and connect the strongest observation to a test the business can actually evaluate.
Co-founder and CEO of Stackmatix, startup advisor, and former Head of Sales at MightyHive. Connect with Matt on LinkedIn ↗ · More about Matt →