Quick answer: qualifying leads from ads means deciding, before anyone spends an hour on follow up, whether an enquiry is worth that hour. Use two axes. Fit asks whether this person belongs to the kind of business you can serve; intent asks whether they are ready to buy now. Score both, set a cut off, write it down. Measuring marketing qualified leads then means counting how many enquiries clear that cut off, what each cost, and what share sales accepts. Without a written cut off you do not have an MQL count. You have a form fill count with a better name.
This page answers one question: how to qualify the leads paid campaigns produce, and how to measure the qualified ones. Generating them is a separate job, covered in lead generation.
Lead, MQL, SQL: what each word has to mean before it can be counted
These three words cause more wasted meetings than any others in Malaysian marketing teams. Rarely because people disagree about the definitions: because nobody wrote any down, so marketing and sales use each term for something different and then argue about the number.
| Stage | Who decides | What it means | Write it down as |
|---|---|---|---|
| Lead | Nobody: it is a count | Any enquiry with contact details, from any source | “A form submission, WhatsApp chat or call with a name and a way to reply” |
| MQL (marketing qualified lead) | Marketing, from recorded signals | A lead that clears the agreed fit and intent cut off | “Fit score 6 or more and intent score 5 or more” |
| SAL (sales accepted lead) | Sales, within one working day | An MQL sales agrees to work, or rejects with a reason | “Accepted, or rejected with one reason from a fixed list” |
| SQL (sales qualified lead) | Sales, after a conversation | A confirmed opportunity with a need, a budget, a decision path and a timeline | “Quotation requested or site visit booked” |
| Customer (closed won) | The invoice | The opportunity became paid work | “First invoice issued” |
Two things follow. An MQL is whatever your business says it is, which is why borrowed benchmarks are close to worthless. And the SAL step, the one most SMEs skip, is what stops the argument: when sales must accept or reject each MQL within a working day and give a reason, you get a weekly list of rejection reasons, and that list is your targeting brief.
These five stages sit inside the wider digital marketing funnel, and the nurture track that catches everything below the cut off belongs to your marketing automation setup.
A vocabulary warning: in Malaysia, searching “MQL” mostly returns MQL5, the MetaTrader language used by forex traders. Spell out “marketing qualified lead” at least once in documents and dashboard fields.
Why leads from ads need different qualification from organic leads
Someone who searched “waterproofing contractor Shah Alam” and read two pages before enquiring has done part of the qualifying work themselves. Someone who saw a Facebook ad while scrolling at 11pm has not. Both arrive in the same inbox, looking identical.
| Organic search | Google Search ads | Facebook and Instagram ads | |
|---|---|---|---|
| What the person was doing | Looking for a solution, then reading your pages | Looking for a solution, clicked an ad | Scrolling, not looking for anything |
| Intent at the moment of enquiry | Usually high | High to medium, depending on the keyword | Mixed: some ready now, many curious |
| Spread of lead quality | Narrow | Moderate | Wide, by design |
| What to qualify on first | Fit | Fit, then the specificity of the request | Intent: timeline and what they actually asked for |
So “our Facebook leads are rubbish” is usually a misdiagnosis. Social ads are meant to produce a wider spread of quality than search. The work is to sort them cheaply and tell the platform which ones were good, so the next batch improves. Channel costs are covered in our guides to Facebook ads cost in Malaysia and Google Ads pricing.
The two axes: fit and intent
Most qualification frameworks are sales frameworks built for outbound calling. BANT, which came out of IBM, scores budget, authority, need and timing. That works when a salesperson is asking the questions. It works badly on a 60 second old form fill, because inbound enquirers rarely volunteer budget or authority, and scoring those as zero punishes a good lead for being new.
For paid traffic, split the question in two. Fit is about the account and does not change this week: industry, size, location, whether you can serve them. Intent is about the moment and decays fast: what they asked for, how specific it was, which page they enquired from, whether they named a date.
| Fit signal (out of 10) | Points |
|---|---|
| Located in an area you serve | 3 |
| Asks for a service or product you actually sell | 3 |
| Business type or sector you serve well | 2 |
| Job or company size within your usual range | 2 |
| Intent signal (out of 10) | Points |
|---|---|
| Names a timeline within the next three months | 3 |
| Describes a specific need, not a general question | 3 |
| Enquired from a service, pricing or quote page rather than a blog post | 2 |
| Returning visitor, or a second enquiry within 30 days | 1 |
| Gave a phone number or WhatsApp and asked to be contacted | 1 |
Then set the cut off. A workable starting rule for an SME with one or two salespeople:
| Score | Verdict | What happens next |
|---|---|---|
| Fit 6+ and intent 5+ | MQL | Sales replies within one working hour, and accepts or rejects within one working day |
| Fit 6+ and intent below 5 | Nurture | Helpful reply, then into the nurture sequence; rescore if they enquire again |
| Fit below 6 and intent 5+ | Manual review | Marketing checks it the same day: refer, decline politely or pass to sales |
| Fit below 6 and intent below 5 | Not qualified | Polite reply, no sales time; note the reason for targeting review |
These weights are a starting point. Run them for six weeks, compare the scores against what sales closed, and move the numbers. A scoring model never revised against outcomes is decoration.
Asking the qualifying questions without killing your volume
Every field you add removes people who would otherwise have submitted. That is not a reason to ask nothing, but a reason to know what each field earns you.
| Field | What it earns you | Cost in volume |
|---|---|---|
| Name and phone or WhatsApp | A way to reply quickly | None: you need it |
| A written channel and a record for the CRM | Low | |
| State or service area | The strongest single fit signal | Low as a dropdown |
| Service needed | Fit, and routing to the right person | Low as a dropdown |
| Timeline | The strongest single intent signal | Low as a dropdown |
| Budget band | Filters out mismatched expectations | Noticeable: add it only when volume exceeds calling capacity |
| Company size or headcount | Fit, for B2B offers | Moderate, and often left unread |
| Hidden: page URL and campaign identifiers | Source attribution all the way to the sale | None: the visitor never sees it |
Four or five fields plus an optional message is the sweet spot. Getting more enquiries than you can call? Add the budget band. Getting too few? Strip back to name, contact and timeline and qualify on the call. More on form behaviour in our notes on improving website conversion rate and landing page optimisation.
WhatsApp is the awkward one, because a large share of Malaysian enquiries arrive there and a chat window has no fields. Two fixes. Use a pre filled first message that differs by page, so the opening text identifies the offer. Then make your first reply a two question qualifier phrased as a human would phrase it: “Happy to help. Two quick things so I send the right numbers: which state are you in, and when are you hoping to start?” See WhatsApp Business API in Malaysia for doing this at volume.
What the PDPA lets you ask and keep
The Personal Data Protection (Amendment) Act 2024 raised the stakes for sloppy lead handling. Penalties for unlawful processing reach RM1,000,000, a data protection officer is mandatory above the stated thresholds, and breach notification to the Commissioner is required. For qualification specifically:
- Collect for a stated purpose. “We will contact you about your enquiry” does not license a newsletter subscription.
- Ask only what you will use. A headcount field nobody reads is liability held for nothing.
- Keep the consent record with the lead: the time, and the wording shown.
- Give people a working way to be removed, and act on it.
- Ad platform uploads count. Sending a hashed email back to Meta or Google is processing.
None of this stops you qualifying. It stops you hoarding.
Measuring marketing qualified leads
Six numbers. Everything else is decoration.
| Metric | How to calculate it | What it tells you |
|---|---|---|
| Leads | Count of enquiries, by source and campaign | Volume, and nothing else |
| MQL rate | MQLs ÷ leads | Whether targeting and the offer attract the right people |
| Cost per MQL | Ad spend ÷ MQLs | The price of a lead worth calling, the figure to manage budget on |
| Sales acceptance rate | SALs ÷ MQLs | Whether the MQL definition matches what sales will actually work |
| MQL to closed won rate | Customers ÷ MQLs, for leads from the same period | Whether qualified leads turn into revenue |
| Speed to first reply | Median time from enquiry to first human reply | Whether good leads are going cold before anyone answers |
Report them by source, not only in total. A campaign with a high cost per enquiry and a high MQL rate usually beats a cheap one with a low rate, and only the split shows it. These sit underneath the campaign level figures in our guides to SEM metrics and ROAS, which measure spend efficiency rather than lead quality.
Two traps. Attribute at the point of the enquiry, not the sale, and keep the campaign identifier on the lead record all the way through the pipeline: if your CRM loses the source at quotation stage, every number below cost per MQL is a guess. And agree a reporting lag, because with a six week sales cycle this month’s MQL to closed won rate describes leads from six weeks ago, and comparing it to this month’s spend misleads every time. Our companion piece on tracking forms and WhatsApp enquiries through to sales covers the plumbing.
Why published MQL to SQL benchmarks will not help you
Look up the average MQL to SQL rate and you will find figures between roughly 13% and 40%, all stated with equal confidence. First Page Sage, using its own client data, puts B2B SaaS near 13%. A survey of 939 B2B companies published by Optifai reports about 40%. Both are vendor datasets and the gap between them is not noise, it is the definition: count every whitepaper download as an MQL and your conversion rate looks terrible, count only pre screened leads and it looks excellent.
So benchmark against yourself. Measure your own rate across one full sales cycle, write it down, and beat it. For what paid channels cost in this market, our digital marketing benchmarks page is more use than imported lead ratios.
Telling the ad platforms which leads were good
This step turns qualification from internal tidying into better ad performance. Both platforms can optimise towards qualified leads rather than form fills, but only if you feed the outcome back.
Google Ads. Import the qualified or won outcome as an offline conversion, keyed to the click identifier captured with the lead, and value the qualified event above the raw lead so bidding chases the event you care about. Click identifiers have a limited life, so the import must run regularly, not quarterly.
Meta. Conversion Leads optimisation does the same job for Facebook and Instagram lead ads. Meta’s own developer documentation sets out the requirements, and three of them catch people out:
| Requirement | What it means in practice |
|---|---|
| The Meta lead ID travels with every update | Your CRM must store the lead ID from the Instant Form and send it back with each stage change. Matching on email or phone alone is not enough. |
| The events carry specific parameters | CRM stage events are sent through the Conversions API marked as system generated, with the event source set to CRM and the name of your CRM included, as the documentation specifies. |
| It works only with Instant Forms | Leads must come from Meta’s native lead forms. Traffic sent to a form on your own website does not qualify. |
That last row matters. If your Meta campaigns send people to a form on your own website instead of an Instant Form, Conversion Leads is not the tool and you are back to standard site conversions. You will also read that you need 200 leads a month or 50 conversion events a week first. Those figures come from agencies and partners, not Meta’s documentation, which states no minimum. Treat them as rough guidance on when the model has enough data.
Google lead form assets change where qualification happens: the form sits inside the ad, the person never reaches your landing page, and your only qualifying fields are the ones Google offers. Eligibility needs a clean policy history, an eligible vertical and a published privacy policy, with spend thresholds and advertiser verification applying to some campaign types. Those thresholds moved during 2026, so check the help page before planning around them. The detail that bites: leads can only be downloaded for the last 30 days and Google keeps them 60, so a manual CSV routine will eventually lose you leads. Connect a webhook or the API to your CRM.
A 30 day plan
| Week | What to do |
|---|---|
| Week 1 | Write the lead, MQL, SAL and SQL definitions and get sales to sign them off. Build the fit and intent scoring sheet. Add a timeline field and hidden page URL and campaign identifier fields to every form. |
| Week 2 | Score every new lead. Start the SAL rule: sales accepts or rejects each MQL within one working day, choosing a reason from a fixed list. Set up the two question WhatsApp first reply. |
| Week 3 | Build the weekly report of the six metrics, split by source and campaign. Set up the Google Ads offline conversion import and, if you run Instant Forms, Meta Conversion Leads. |
| Week 4 | Review the rejection reasons and adjust targeting and form fields. Send the first qualified outcomes back to the platforms. Note what to change in the scoring weights at the six week review. |
Mistakes that cost the most
- Reporting cost per lead upward while qualifying internally. The two numbers drift apart and the budget gets cut on the wrong one.
- Letting sales reject MQLs silently. No reason logged, no learning.
- Scoring budget and authority on an inbound form, which hides good leads for being new.
- Changing the MQL definition without restating the history. The trend line dies and nobody notices for a quarter.
- Adding a seventh field to improve quality, then wondering why volume halved.
- Qualifying hard but replying in two days. First reply speed moves outcomes more than scoring refinements do.
Frequently asked questions
What counts as a marketing qualified lead?
An enquiry that meets a fit and intent threshold marketing and sales agreed in writing. There is no universal definition, which is why two companies in one industry report wildly different MQL counts from identical traffic. The useful test: would a salesperson thank you for the handover? If not, it is not an MQL, whatever the dashboard says.
How do I measure MQLs if I have no CRM?
A spreadsheet works under roughly 100 enquiries a month: one row per enquiry, with date, source, campaign, fit score, intent score, verdict, sales decision, reason and outcome. The discipline matters more than the tool. The catch is that nobody updates a spreadsheet reliably past the second week, which is where a CRM starts paying for itself.
What is the difference between an MQL and an SQL?
Marketing decides the MQL from recorded signals, before anyone has spoken to the person. Sales decides the SQL after a conversation that confirms a real opportunity with a budget, a decision path and a timeline. The gap between the two counts is your handover quality, and a wide gap usually means the MQL definition is too loose rather than that sales is being difficult.
Should I qualify leads before or after the first call?
Both, at different depths. Score automatically before the call to decide who gets called first and who goes to nurture. Qualify properly on the call, because that is where budget, authority and the real decision process surface. Pre call scoring is triage, not diagnosis, and treating it as diagnosis is how good leads get filed away.
Will adding qualifying questions reduce my leads?
Yes, and that is sometimes the goal. Expect fewer submissions with each added field. The question is whether the enquiries you lose were ones you wanted. If your team cannot call everyone within a day, trading volume for quality is an improvement. With spare calling capacity, keep the form short and qualify on the phone.
How do I qualify WhatsApp enquiries from ads?
Use a different pre filled message per ad or page so the opening text identifies the source, then ask two qualifying questions in your first reply, phrased conversationally, and log both answers against the contact. Avoid long automated question trees: drop off is high and they read like a form, which is what people chose WhatsApp to avoid.
Can Google and Meta optimise for qualified leads rather than form fills?
Yes, if you send the outcome back. On Google, import the qualified or won event as an offline conversion with a higher value than the raw lead. On Meta, use Conversion Leads optimisation, which needs the lead ID and the right event parameters and works only with native Instant Forms. Both need the feed on a regular schedule.
The bottom line
Qualification is not a filter you buy. It is a definition you write and then revise against outcomes. Write the MQL definition, add a timeline field and a page URL field, make sales accept or reject within a day with a reason, keep the source attached to the lead all the way to the invoice, and feed the result back to the platforms. Six metrics, one page, weekly.
If you want a second pair of eyes on how your paid campaigns are being measured, check your AI Visibility & Readiness Audit free with us.
