Brand promotion: this article introduces Guangsuan's Google Ads management service. It is a vendor introduction, not an independent review, a customer story or a commissioned evaluation.
A health analytics platform that sells to employers and health plans can have everything apparently working. The demo form submits, the calendar fills, the ads account reports conversions, and the sales team still says the pipeline is thin. Both statements can be true at once, because they count different events at different stages. A demo request is a site or product event. A qualified opportunity is a judgement about the buyer's covered population, data access, security review and timeline, and most of that judgement happens where the ad account cannot see it.
This article answers one question: when demo requests arrive but qualified pipeline does not grow, what should a Google Ads management engagement cover beyond campaign setup and bid changes, and in what order should the parts be checked? It is written for growth and demand teams at B2B health analytics or population-health platforms that already run search campaigns. The checks fit search campaigns aimed at buyers who are actively looking for a vendor. For awareness or reach formats, only part of the list applies.
Short answer
Separate five records that are often reported as one number: the action the account counts as a conversion; the searches the budget actually bought; the page that received the click; the definition sales applies before calling a request qualified; and the route the request takes after submission. Check them in that order. Each one is cheap to inspect, and each one narrows the next question.
Two conditions belong with that answer. Write the scope down, because "management" can mean weekly bid adjustments or the whole path from search term to sales feedback. And treat a fixed number of qualified leads as something no provider can promise. Guangsuan states that it does not promise a fixed inquiry volume or deal outcome. What can be agreed is the work, the review cycle and the data used to judge it.
| Count | What it shows | What it cannot show |
|---|---|---|
| Click | Someone reached the site from an ad | Whether the page answered the query |
| Conversion | An action the account was configured and able to record, such as a form submission | Whether the organisation could buy, or what the request was worth |
| Demo request received | A person asked for a conversation and the message reached an owner | Whether the use case, population or data access fit |
| Qualified opportunity | A sales judgement that a written condition was met | Whether a deal will close, or which query deserves credit |
| Closed business | An outcome recorded in your own systems | Anything about the other requests in the same period |
Guangsuan's guidance keeps these stages apart rather than folding them into one performance figure: the contact-button click, the form that arrives on the site, the inquiry sales accepts, and the contract or payment stage. That separation is the starting point for the checks below.
1. Agree what the account is counting
Open the conversion action list and write down, for each action, what actually triggers it: a click on a contact button, a thank-you page after a form, or a server-side event. Record which actions are primary and which are secondary, because that choice decides what reporting shows and what bidding optimises toward. Note the counting setting, the attribution model, and whether the report uses click time or conversion time.
A health analytics account tends to accumulate more actions than the team remembers: demo request, content download, webinar registration, newsletter, a "talk to sales" click on a pricing page. If several of them are recorded as the same conversion, the report can be accurate and still unable to tell you which one moved. Guangsuan's published scope places analytics linking, measurement and conversion setup inside the paid work rather than beside it.
Google's own help draws a related line: a qualified lead is assessed further inside a lead system, while a converted lead has completed a step the business defined, which is not automatically a customer. Guangsuan's guide to returning qualified lead feedback cites that distinction. It is the reason a demo request and a qualified opportunity should not share a column.
2. Read the searches the budget actually bought
The search terms report shows the queries that triggered the ads. It is the only place where the intent behind a paid click is visible, and it is incomplete: Guangsuan's guidance notes that queries with insufficient activity can be omitted under data privacy standards, so the list is a sample large enough to act on, not an inventory.
A first pass can sort the visible queries into three groups.
- Buying. A use-case, capability or supplier query, or a comparison between vendors. Keep it, and check that the landing page answers it.
- Unresolved. Definition and methodology questions, integration questions, and price questions from someone who has not described a covered population or a timeline. Watch under a written condition, or route to a page that can turn research into a conversation.
- Not a buyer. Job searches, student and course requests, free-template downloads, and vendor pitches. Exclude at the narrowest workable scope, with the business reason recorded.
For this category the exclusions are easy to get wrong, because one word can belong to two very different searchers. "Risk score" may be a buyer comparing approaches or a student looking for a formula. "Population health" may be a health plan evaluating vendors or a clinician looking for a definition. The decision is commercial, not lexical.
Keyword research and negative keyword maintenance are separate jobs that pull in opposite directions: one widens the set of searches the account can match, the other removes demand the business cannot serve. A month spent only on the first quietly broadens the account. Each exclusion should leave a record: the query, the negative used, its scope and the reason, so the decision survives a change of account manager.
3. Check whether the page answers the promise the ad made
A click is a question. If the ad names a capability and the page opens with a company introduction, the visitor has to hunt for the answer the ad already promised.
For a health analytics platform the questions are concrete: what the product does and does not do; which data sources it uses and what happens when one is missing; whether the output is a score, a cohort or a report; who inside the buyer's organisation uses it; how deployment and security review work; and what the demo request actually commits the buyer to. A page that answers four of six still leaves the buyer to ask sales.
Guangsuan describes the intended route as matching the query to the relevant product or service page rather than routing every click to the homepage, and places landing page analysis, optimisation and URL path matching inside the paid scope. That placement matters when comparing proposals. If a plan stops at keywords and bidding, the page where the promise is kept or broken has no owner.
The demo page is part of the same check. If it does not say who joins the call, how long it runs, what the buyer should prepare and what happens afterwards, the request arrives with expectations the sales team has to repair.
4. Write a definition sales can apply the same way twice
"The leads are poor" is not a definition. A working definition for this buyer type might include four conditions: the organisation is an employer, health plan, provider network or consultancy the business can serve; a population or covered-lives figure in a serviceable range is available; the relevant data sources and security review path exist; and a named person can continue the conversation within a stated period. It is a starting point to adapt, not a template to copy, because a self-insured employer and a national health plan need different confirming evidence. Qualification criteria should be adapted rather than copied between businesses for that reason.
Then make the record usable. A short status list is enough: submitted, contacted, awaiting information, qualified, not a fit, business outcome, each with the time it happened and the person who decided. Ask sales to record a specific reason for "not a fit" rather than a single label, because a routing delay and a genuinely unsuitable request look identical when both are marked invalid. Guangsuan's guidance on reviewing inquiries asks for exactly that specificity instead of one blanket verdict.
Two people should be able to apply the definition to the same request and reach the same answer. If they cannot, the account will keep optimising toward a number that sales does not recognise.
5. Follow the record from the form to the sales stage
A submission that displays a success message has not necessarily reached a person. Someone has to receive it, own the follow-up, and record what happened. Give each request an internal ID that travels from the form into the sales record. Without it, the ad report and the CRM can only be compared as totals, and a total that mixes organic, referral and paid contacts will never explain the campaign.
If later stages are imported back to the ad platform, check the route and the rules before assuming it works. Guangsuan's guide explains that hashed customer data in enhanced conversions is matched to site and signed-in data for attribution. It does not judge whether a lead fits, and it cannot backfill identifiers that were never collected.
There is a boundary that matters more than usual here. Guangsuan's guidance notes that enhanced conversions cannot be used to measure conversions in sensitive categories such as health or medical information, and that consent or hashed data does not remove that limit. A health analytics vendor should therefore confirm with its privacy or legal lead which fields, if any, may be used for ad measurement, instead of assuming that everything a demo form collects is fair game. Keep identifiable health information out of shared reports and error logs regardless.
Finally, align the period before comparing the two systems: same date range, time zone, conversion action and attribution window. A comparison that mixes click-time and conversion-time reporting can show a difference that has nothing to do with lead quality.
A hypothetical example (invented numbers)
The figures below are teaching numbers for a fictional account. They are not a client result, a benchmark or an observed outcome.
| Stage in the path | Count in the example | What the next question becomes |
|---|---|---|
| Demo requests recorded in the account | 48 | Is the account counting the action the business actually cares about? |
| Requests sales contacted | 39 | Did the request reach an owner, and how quickly? |
| Requests that met the written definition | 14 | If most submissions never qualify, check the queries and the page rather than the bid strategy. |
| Qualified requests still active after 60 days | 9 | Follow-up and sales cycle, not the ad platform. |
None of these figures is wrong on its own. Treating all 48 as qualified would replace the account's picture of a good lead with the picture of a form fill, and the requests that qualify later belong in a later comparison rather than in this month.
What better measurement does not fix
Improving the definition changes what the account optimises toward; it does not create buyers. If demo requests come from organisations outside the serviceable market, or the product does not yet answer the problem the ad describes, the gap is in targeting or positioning, and a rebuilt conversion action will describe the mismatch more precisely rather than remove it. Guangsuan's page separates three possibilities before any budget increase: the traffic is not the right audience, the page is not holding the visit, or the data is not clear enough to tell which. Only the first is repaired inside the ad account.
Where Guangsuan fits
Guangsuan (光算) provides Google Ads management for export and B2B inquiry generation. Its published description covers search terms, ads, landing pages and conversion tracking, with the stated aim of optimising around real purchasing demand rather than clicks alone.
The published scope is a set of work items combined per project: industry research, promotion strategy and account diagnosis; account structure and campaign setup; budget, targeting, ranking and bid optimisation; keyword research; search term report analysis and negative keyword maintenance; landing page analysis, optimisation and URL path matching; analytics linking, measurement and conversion setup; creative optimisation; conversion performance, deal feedback and ROI analysis; and periodic reporting with review meetings. Material production, page development and data integration are confirmed in the plan and contract rather than assumed.
The stated basis for ongoing optimisation is three inputs read separately: Google Ads account data, website analytics and sales feedback. That is a useful expectation for a health analytics team, because a review that reads only the ad account will keep treating page problems and follow-up problems as bidding problems. Guangsuan also states that it does not promise a fixed inquiry volume or deal outcome, and that it was founded in 2019 and works on a corporate contract with a defined delivery scope.
If the immediate argument is about conversions in the report and nothing in the CRM, Guangsuan's guide to returning qualified lead feedback from Google Ads goes deeper into the record side. The service scope, fee structure and review process are published on the Guangsuan Google Ads management page.
FAQ
Is a demo request the same as a qualified lead?
No. A demo request is an event the site or product recorded. A qualified lead reflects a judgement about the organisation, the use case, the population in scope and the ability to continue, which the ad account does not hold. Both can be counted accurately and still differ.
Should we switch campaign goals to qualified leads immediately?
Only when the definition is written, applied consistently and the volume is high enough to learn from. If qualified requests are few or arrive months later, replacing every campaign goal at once removes the signal used to compare periods. Agree which campaigns change, how long to observe them, and what would justify reversing the change.
Can uploading CRM data close the gap?
It can improve attribution when the record path and the platform rules allow it, and it does not create fit. Sensitive categories such as health or medical information remain excluded from enhanced conversions for measurement, and a technical upload route does not settle whether the underlying use is permitted.
Do we need a CRM before we can start?
No. A controlled spreadsheet with a request ID, a status, the time it happened and an owner is enough to begin, provided personal and health data stay in an access-controlled system. A tool does not create the definition; the definition is what makes the tool useful.
What to bring to a first conversation
- The website address and the pages that already receive paid traffic.
- The main products, the buyer types you can serve, and the countries or languages you can support.
- A budget range and the period over which you expect to review progress.
- The current conversion action list, with what triggers each one.
- The definition sales uses now, and the reasons recorded for requests marked not a fit.
- The questions sales answers most often, which tend to become the sections both the ad and the page should point at.
- Historical account data you are authorised to share, used to assess structure, search terms, conversion setup and budget performance.
The order above is deliberately unglamorous. Fixing the definition without reading the queries, or reading the queries without checking the page, produces a tidier report and the same conversation next quarter.