Without “harm” or “scam”: finding questions worth checking in Google Suggest
How to find landing-page topics worth checking in neutral search queries — without inventing audience “pain points”.
1. What search suggestions add to audience research
Before launching a landing page, campaign or feature, an editor tries to answer: what questions might arise before someone acts, and which ones does the page still leave unanswered?
Interviews, reviews, support conversations, web analytics and keyword research all help. They complement one another; each shows a different part of the picture. An interview usually describes an experience after the fact. A review often appears after a notably good or bad outcome. Search-volume data can help estimate scale; autocomplete provides a different view by surfacing related query wording.
Google Suggest provides another source of search language. A suggestion shows that Google's autocomplete system surfaced that wording at the time of collection. It does not tell us how often the phrase was searched or why Google selected it.
For this experiment, we did not append
harm, scam, fraud, dangerous or disadvantages to the root queries. Instead, we collected suggestions to two levels, passed them to a model and then checked which concerns were actually present in the wording and which were added by the model.It does not produce a complete picture of the audience. It can turn a broad topic into concrete questions to explore in interviews, an FAQ or a page review.
The test set: 235 English root queries without deliberately added negative modifiers, 10,569 unique level-1 and level-2 Google Suggest phrases, and one Gemini 3.8 Flash answer for each root.
2. Protocol: from a root query to a hypothesis
How the chain was built
Code:
root query: “mri with contrast”
│
├── level 1: “mri with contrast side effects”
│ │
│ ├── level 2: “mri with contrast side effects next day”
│ └── level 2: “brain mri with contrast side effects”
│
└── level 1: “mri with contrast vs without”The first level often contains brands, locations, prices or basic informational terms. The second level may add a question, restriction, post-purchase problem or specific concern. Sometimes the second level adds nothing useful. That is a result too.
Collection and review protocol
- Collection and preparation in A-Parser: for each root query, A-Parser fetched Google Suggest results, expanded them to two levels, removed exact duplicates and assembled a single list for analysis.
- Seed queries: 235 English root queries without deliberately added negative words.
- Why 235: the 235 roots are the deduplicated English queries present in the saved dataset used for this experiment. They cover a mix of consumer, business, software, education, financial and health-related commercial topics; this is a convenience dataset, not a representative sample of the market.
- Depth: a root query is level 0; only suggestions from levels 1 and 2 were included in the final set. Level-3 results from the depth-three tests were excluded.
- Unit of analysis: an exact text suggestion. Whitespace was normalized, and duplicates within the English set were removed.
- AI step: A-Parser sent the collected suggestion list and an English analysis prompt to Gemini 3.8 Flash through OpenRouter. The model was asked to identify only possible concerns, doubts or selection barriers based on the words in the list and to return the answer in English.
Gemini received the full suggestion set for each root but was instructed to return only a short list of candidate issues.
- Model output: the consolidated dataset contains one retained model response for each root query.
- Review of conclusions: a model can complete successfully and still attribute a meaning that the phrase does not contain. Before publication, an editor manually checked the examples used here, separating literal suggestion wording from added interpretation.
When several saved responses existed for the same root, the dataset builder preferred one with candidate issues over a no-signal response, then applied production recency and deterministic tie-breakers. The 149/86 figures therefore describe the retained dataset, not independent repeated model calls.
Gemini is a fast first-pass reader, not a substitute for audience research. An editor still needs to verify every candidate issue.
3. What the test collected
The collected suggestions do not provide a complete picture of the audience. They show how broad topics break down into concrete search wording: from navigation and pricing to questions about terms, restrictions and outcomes.
| Metric | Test result |
|---|---|
| Root queries | 235 |
| Level-1 and level-2 suggestions | 10,569 |
| Roots flagged by Gemini for manual review | 149 |
| Roots where Gemini returned no candidate issue | 86 |
149 is a Gemini-generated review queue, not a count of confirmed problems or a market diagnosis.
For 149 root queries, Gemini returned at least one candidate issue based on the suggestion list. This is a model classification for editorial review; it does not tell us how common the question is or whether it changes a prospective customer's decision.
The process was:
4. Examples: what suggestions can reveal
At the first level, suggestions commonly refine a choice by brand, location, price, model or service type. At level two, suggestions sometimes become much more specific: a restriction, a comparison, a side effect, or a problem the searcher is trying to understand.
The six examples below use English roots, suggestion lists and their corresponding model responses from the test set. Each card moves from the observed wording to the model's interpretation and then to a possible page action.
With
locale=en, A-Parser sends the analysis instructions in English and asks Gemini to return a single-line response in English. The interpretations below are not independent evidence of user motivation.Before publication, an editor checks claims with the sales team, the product team and, when necessary, a subject-matter specialist. A suggestion can point to a page topic; it does not justify a claim the service cannot support.
The example suggestions below are illustrative excerpts; Gemini received the full suggestion list for each root.
Laser hair removal: home use and suitability
What the data shows:
What may be missing from the page: A page built around speed, price or a device feature list can leave the practical fit unclear. Prospective buyers need to know which skin and hair types the treatment is intended for, what limitations apply, when home use differs from professional treatment, and which questions require clinical advice.
What to do: Make suitability and safety a central part of the page. Cite the relevant limitations and selection criteria, then give readers a clear path to professional advice. Medical claims require expert review before publication.
Cloud accounting software: migration and setup complexity
What the data shows:
What may be missing from the page: Integrations and dashboards do not answer the adoption question on their own. The page could explain what adoption involves, including initial setup, data migration where relevant, what remains under the customer's control and the situations in which the product is not a good fit.
What to do: A migration checklist and a first-week setup plan would make the process concrete. Include a plain-language comparison of cloud and non-cloud deployment options.
Business VPN service: need, regional access and configuration
What the data shows:
What may be missing from the page: Generic security claims leave several buying questions open: when a business needs a VPN, whether the service works in China, and how the deployment options differ. Regional availability also needs to be documented.
What to do: Give readers a “Do we need this?” decision guide and explain the implementation choices. Document regional availability clearly and show when the information was last updated.
Business email hosting: reliability, limits and included features
What the data shows:
What may be missing from the page: Mailbox size, storage and price answer only part of the buying question. A business also needs to know what happens when email stops working, which limits apply to each plan, whether the service is a good fit and what the platform includes.
What to do: Spell out deliverability, availability, plan limits, included features and support paths. A clear comparison of the available business email options would make the choice easier.
Crypto exchange: legal questions and identity verification
What the data shows:
What may be missing from the page: A registration flow without clear access conditions can leave important questions unanswered. Visitors need to see when verification is required, what information is requested, which terms vary by jurisdiction and how to check current eligibility.
What to do: Set out the onboarding process and identity checks in plain language, with supported access scenarios and jurisdiction-specific terms. Searches for no-verification access should not be presented as a promise to bypass rules.
Business cloud storage: security, reliability and scalability
What the data shows:
What may be missing from the page: Capacity and price are only part of the decision. The page also needs to cover data protection, reliability expectations, the trade-off between cloud storage and existing enterprise infrastructure, and how capacity and limits change as the business grows.
What to do: Use the page to document protection and availability, capacity tiers, known limits, scaling options and the migration path from existing infrastructure. Make the features and limits of the current plan easy to find.
5. Where a suggestion ends and a guess begins
A Gemini response is an interpretation, not a conclusion about a particular searcher. Even when the model relies on the collected suggestions, it can add a motive, emotion or problem that the wording itself does not contain.
For example,
vpn service is unavailable cisco points to a service-availability issue in a named client. It does not tell us why a particular searcher needs a VPN or what they intend to do with it.Before publication, an editor should do three things:
- Keep anything explicitly present in a suggestion.
- Check any proposed reason for the search through sales, support, interviews or product data.
- Reject fears, intentions or problems we have no evidence to associate with the searcher.
Price, reviews, free, how to choose, a location, a model or a platform may all be useful topics for a page; none is proof of a pain point.A suggestion helps an editor choose a question to investigate. It does not justify claiming to know what a searcher fears or why they have not decided.
6. Method limits
- Suggestions are not exact search volume. A suggestion shows that Google's autocomplete system surfaced that wording at the time of collection. It does not tell us how often the phrase was searched, how many people searched it or why Google selected it.
- One collection does not replace market research. It cannot support claims that “most users are afraid” or that a market has a particular motive.
- Suggestions are dynamic. Results depend on language, region, time, the starting wording and Google's suggestion system.
- No suggestion does not mean no question. The root may be too narrow, the wording may be poor, or the query may not have produced more specific suggestions at the chosen depth.
- AI output is interpretation. Even when a model supplies supporting phrases, it can overstate what they show. The dataset is raw material for verification, not a complete picture of the audience.
7. Conclusion
Google Suggest is therefore a starting point for research, not a measurement of demand or a window into what “the market really fears.” It can still highlight questions a page may not yet answer.
The useful outcome is not a list of dramatic “pain points,” but several checkable topics: what to explain, which conditions to state clearly and which scenarios to make clearer before someone takes action.
If the exercise leaves an editor with a short list of topics for a page, an interview or a product decision rather than unsupported assumptions about user motivation, it has done its job.
In this form, AI does not replace audience research. It helps decide what to research next.
Appendix: experiment materials
Below is a ready-to-import preset so you can repeat the experiment without building the scraper manually.
Code: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How to run it:
- Open “Task Editor” → “Task” → “Import preset” and paste the code from the spoiler.
- The import creates two tasks: Google Suggest + Gemini (RU) and Google Suggest + Gemini (EN). The
JS::GoogleSuggestGeminiTypeScript scraper is already included in the preset. - Replace the test root query in the “Query list” field with your own list. The preset already uses
rufor the Russian task andenfor the English task. - Enter your key in the parser's OpenRouter API key setting or override. The key is entered directly in the A-Parser interface.
- Run the task. The result is saved to TXT, one record per line:
root query<TAB>model response.
Last edited: