Financial Services PPC Case Study in Pakistan

The Client

A Lahore-headquartered licensed non-banking finance company offering Shariah-compliant financing to small and medium enterprises across the Punjab industrial belts — Lahore, Faisalabad, Gujranwala, and Sialkot. The portfolio had three products: a murabaha working-capital facility from PKR 500k to PKR 25M, an ijarah equipment-financing line for machinery and commercial vehicles, and an invoice-purchase facility advancing up to 80% of verified invoice value. The average sanctioned facility ran near PKR 2.8M, and the company’s credit appetite sat with established manufacturers and traders — two years of trading history and monthly bank inflows above a defined floor.

Distribution was relationship-officer led: six officers covering Punjab, fed by a two-person marketing team running roughly PKR 750k a month of in-house Google Search through one campaign, one generic landing page, and a short enquiry form. Every submission was called within a business day, documents requested, and the file moved to a credit committee. It is a structure common to Pakistani NBFCs: light headcount, conservative credit culture, and a marketing function that owns spend but not the funnel beyond the form fill.

The engagement began when the company’s own compliance function froze the flagship ad claim — “48-hour financing” — after an internal review found the median decision time on complete files was four working days, not two. Marketing had paused most testing rather than argue, disapprovals from the platform were accumulating, and the head of distribution approached WeProms Digital for PPC management services with one condition attached: nothing we shipped would put the license, the Shariah board, or the company’s name at risk. This page is an illustrative composite — the framework, economics, and constraints are drawn from patterns WeProms sees across Pakistani regulated lenders, applied to a representative engagement.

The Problem

Five issues were converting a reasonable monthly budget into a weak pipeline:

• More than a third of the spend bought ineligible intent. Six months of search-query data showed 38% of click spend landing on queries that could never become SME financing applications — personal-loan variants, “loan without documents” searches, job-seeker queries, and salary-advance intent. Broad match with no negative discipline was quietly donating budget to consumer Google.
• Claims the company could not stand behind. “48-hour financing” and “lowest markup in Pakistan” were the two workhorse headlines. Neither was substantiated; the first was frozen by compliance, the second had no verifiable basis under Competition Commission deceptive-marketing expectations around comparative claims. Fourteen ads had been disapproved over the prior quarter — six for missing financial-product disclosures, four for unsupported claims, two for certification issues, two for destination problems.
• Islamic-structure language used loosely. One ad group described murabaha facilities as “interest-free financing” — phrasing the Shariah board objected to on principle, since a murabaha structure carries a disclosed markup, and presenting it as interest-free conceals the customer’s actual cost. It also flirted with platform policy on misleading financial terms.
• A form that asked nothing and filtered nothing. Name and phone number only. Fifty-four percent of submissions qualified far enough to be worth an officer’s call; the other 46% were consumers, job seekers, or businesses far below the credit floor — each one consuming officer time before being declined.
• Bidding blind to quality. The only conversion flowing back to Google Ads was the form-fill event. Smart bidding had spent months learning to buy cheap form fills, which is precisely the traffic a lender least wants.

Underneath all five sat a process constraint: every creative change queued sequentially through compliance, averaging five days. Testing had effectively stopped, because a test that launches a week late is not a test.

Phase 1 — Compliance and Waste Diagnosis (Weeks 1-2)

Phase 1 produced two documents that every later decision referenced: a query-waste audit and a substantiation ledger.

Search-query forensics. We pulled six months of search-query reports, tagged every query by whether it could plausibly become an application under the company’s credit policy, and bucketed the failures. Thirty-eight percent of spend was ineligible: 21 points on consumer-loan intent, 17 on job, education, and miscellaneous leakage. The audit also surfaced a second-order finding — even the eligible traffic was being sent to one generic page regardless of product, so a manufacturer researching ijarah machinery finance landed on the same copy as a trader asking about working capital.

Query intent class	Share of spend (before)	Share after rebuild
Eligible SME financing intent	62%	93%
Consumer-loan leakage	21%	2%
Jobs, education, miscellaneous	17%	5%

The disapproval post-mortem. We categorized all fourteen prior-quarter disapprovals by root cause. The pattern was consistent: the account was not being rejected for what it sold, it was being rejected for how loosely it described what it sold. Missing markup-and-fee disclosures and unverifiable superlatives accounted for ten of fourteen. This reframed the problem for the client’s compliance head: the fix was not slower review, it was better raw material.

The substantiation workshop. With the compliance lead and a Shariah-board contact in one working session, we went claim by claim. What speed promise could the company stand behind? Trailing-six-month disbursal data showed a median credit decision of four working days on complete files — so “credit decision within five working days on complete files” became the approved claim: specific, verifiable, and safer than the truth it replaced. What could be said about cost? Indicative markup ranges by product and tenor, disclosed on the landing page rather than compressed into ad text. What terminology was off-limits? “Interest-free” for murabaha and ijarah products, “guaranteed” anything, and approval language without the conditional “subject to credit assessment and SECP-regulated criteria.” The output was a one-page terminology guardrail that a copywriter could follow without a meeting — the same discipline that governs digital marketing for banks and licensed lenders in Pakistan, where CCP substantiation expectations, SECP advertisement guidance, and platform financial-products policy overlap.

Baseline frozen. Cost per lead, qualification rate, cost per qualified application, disapproval count, and review turnaround — logged in a shared dashboard before anything changed, so the ninety-day read would be judged against a fixed start.

Phase 2 — Product-aligned Account and Landing Rebuild (Weeks 3-5)

Phase 2 rebuilt where the money went and what happened after the click.

Three campaigns, three products. The single blended campaign became three product campaigns — murabaha working capital, ijarah equipment finance, invoice purchase — each with intent-tiered ad groups (product-specific terms, Shariah-finance-specific terms, and comparison-stage terms), tightened match types, and a shared negative list of roughly 900 queries built from the Phase 1 forensics: consumer-loan vocabulary, job and education intent, “without documents” variants, and adjacent categories the credit policy excluded.

Landing pages that answer an underwriter’s first questions. Each product got its own page: facility range, indicative markup band, tenor options, an eligibility checklist stating the trading-history and inflow floors plainly, and a disclosure block positioned where platform reviewers and the SECP framework expect it. The generic page had converted at 9.5% of sessions because it asked nothing; the product pages converted somewhat lower per session and dramatically better per officer-hour.

The form as first-stage underwriting. The name-and-phone form became a single-screen application asking business name, city, business type, years trading, monthly bank-inflow band, and the product of interest — with a consent checkbox covering the data handling PECA expects. Every field mapped to a real credit criterion, so a submission arriving at a relationship officer was pre-screened against the floor. Form-fill volume dipped for eleven days as the junk traffic starved; qualification rate began climbing immediately.

Closing the measurement loop. Alongside the rebuild we ran the Google Ads conversion tracking setup properly: GCLID captured into the CRM lead record at submission, and offline conversion import configured for two CRM milestones — “qualified application” (the officer confirms credit eligibility and requests documents) and “disbursed.” For the first time, the ad platform could expect to learn which keywords and creatives produced lendable files rather than form fills.

Phase 3 — Claim Library and Qualification Bidding (Weeks 4-8)

Phase 3 turned compliance from a gate into a component system, then aimed bidding at the metric the business actually cared about.

The pre-approved claim library. We assembled twenty-two substantiated claim blocks and four disclosure blocks — every one traceable to disbursal data, the approved markup ranges, or documented process facts — plus the terminology list from Phase 1. Assets assembled exclusively from library components would clear compliance in under twenty-four hours on a fast-lane agreement with the client’s compliance lead; anything novel still took the full review. This is what restored testing velocity: the constraint was never the review itself, it was sending unreviewed raw material into it.

Specificity out-pulled superlatives. The rewritten ads replaced “48-hour financing” with “credit decision in five working days on complete files,” and added concrete targeting statements — “murabaha working capital from PKR 500k to PKR 25M for Punjab manufacturers.” Click-through rate rose from 2.1% to 3.6% on the same impression base. The mechanism is worth naming: specific, verifiable claims self-select the right audience. A business owner who clicks a disclosed facility range arrives half-qualified; a browser who clicks “48-hour financing” arrives with a hope and no documents.

Bidding on qualified applications. Once the offline conversions carried four weeks of data, the high-intent campaigns moved from form-fill target-CPA to qualified-application target-CPA, with keywords pruned by cost per qualified application rather than cost per lead. Within a fortnight the account visibly re-priced: keywords that had looked cheap under form-fill accounting revealed themselves as expensive under qualified-application accounting, and vice versa.

Disapprovals, re-examined. Two disapprovals landed across the ninety days — both destination-URL issues from a staging link escaping into a live ad, both fixed same-week. Zero claim-based rejections, against ten in the prior quarter. The claim library had not made the account cautious; it had made it boring to the policy layer, which is the correct state for a regulated advertiser.

Operational synchronization. Ad scheduling was tightened to business hours and geo-weighted toward the Punjab industrial belts, because a financing enquiry called back a day late is a financing enquiry that signed elsewhere. The ads now promise a five-working-day decision; the funnel was re-tuned so the first call happened inside four hours of submission.

Phase 4 — Measure, Feed Back, and Extend (Weeks 8-12)

The final phase compounded the gains and extended the system’s reach.

Value-based bidding. With enough disbursed events imported, bidding graduated to value weights: a disbursal counted five times a qualified application, reflecting the difference in terminal value. The account began preferring the keyword cohorts that produced sanctioned facilities — smaller in raw lead volume, larger in what the company actually exists to originate.

The weekly sales-media loop. A fifteen-minute standing sync between relationship officers and the media side turned recurring disqualifier patterns into account changes: traders in sectors outside the licensed scope became negative keywords; a recurring confusion between the invoice facility and factoring-as-outsourcing became a landing-page clarification; a credit-floor edge case became an additional form hint. This loop, more than any single optimization, is what held the qualification rate at 60% as volume grew.

The Microsoft Ads pilot. In week nine we cloned the highest-intent campaign structure onto Microsoft Ads, funded by trimming residual Google waste so total spend stayed at PKR 750k. CPCs ran roughly 30% lower on B2B-adjacent queries — a familiar pattern, since desktop Bing skews toward office-hours, at-desk searchers — and by week twelve the pilot was contributing about 11% of qualified applications at a better cost per application than the Google average.

Institutionalizing the compliance fast lane. The monthly claim-library review with compliance and the Shariah board became standing governance: new claims enter only with substantiation attached, and retired claims leave the library before they leave the ads. The five-day review cycle ended the engagement at under twenty-four hours for library assets — a change in throughput, not in standard.

The scaling plan. With cost per qualified application holding near PKR 4,750 through weeks ten to twelve, the company approved a 20% budget increase for the following quarter, targeted at the ijarah campaign where disbursal value per application ran highest — spent into a system now measured on the numbers that matter.

Final Results at 90 Days

Metric	Before	After (90 days)	Change
Cost per qualified application	PKR 7,200	PKR 4,750	−34%
Qualified applications / month	104	158	+52%
Raw form fills / month	193	263	+36%
Lead-to-qualified rate	54%	60%	+6 pts
Search CTR	2.1%	3.6%	+71%
Ineligible query spend share	38%	7%	−31 pts
Policy disapprovals / quarter	14	2	−86%
Compliance review cycle	5 days	<24 hours (library assets)	−80%
Application-to-disbursal rate	8%	11%	+3 pts

The arithmetic is deliberately flat-spend: the same PKR 750k a month bought 193 raw fills at a PKR 3,890 average before, and 263 fills at PKR 2,850 after — 60% of which qualified, yielding 158 usable applications at PKR 4,750 each. Nothing in the table depends on a budget increase; every improvement traces to the interventions above. These figures are illustrative outcomes assembled from common patterns WeProms observes across Pakistani regulated-financial-services marketers, not an audited third-party statement — they exist so a lender’s growth team can sanity-check what a compliance-aware rebuild should plausibly return on an account of this shape.

What Made This Work

1. Substantiation solved both problems at once. Replacing “48-hour financing” with the five-working-day claim the data actually supported satisfied compliance, the Shariah board, and platform policy simultaneously — and raised CTR 71%, because SME owners respond to specifics they can plan around. The compliant version was also the better ad.
2. Product-keyword alignment killed the waste before bidding changed anything. Splitting one blended campaign into three product-aligned campaigns with a 900-query negative list removed the 38% ineligible spend outright. That single structural move funded everything that followed.
3. The form did first-stage underwriting. Seven fields mapping to real credit criteria filtered the 46% dead leads before an officer touched them — the cheapest lead-quality intervention available to any lender, and the one that raised qualification rate without suppressing qualified volume.
4. Offline conversions repointed the machine. Until qualified-application and disbursed events flowed back, smart bidding was optimizing for the one outcome a lender least values. Importing CRM milestones is what let the platform spend toward lendable files.
5. The claim library converted compliance into throughput. Pre-approved components with a twenty-four-hour fast lane restored weekly testing inside a regulated environment — without lowering a single standard.

What Teams Can Apply

For Pakistani banks, NBFCs, insurers, and fintechs running paid acquisition:

1. Audit search queries for eligibility before touching bids or ads. Tag six months of queries against your actual credit policy. Whatever share cannot become a customer — consumer intent, jobs, out-of-scope sectors — is budget you are donating, and no bidding strategy recovers it.
2. Replace every superlative with a substantiated operational fact. You likely already own the replacement: median decision times, facility ranges, disclosed markup bands. Regulators, platform policy, and SME buyers all prefer the same thing — verifiable specifics. Run the substantiation workshop once and the copy problem stays solved.
3. Import the CRM milestone you actually care about. Form-fill optimization buys form fills. If your business is built on qualified applications or disbursed facilities, feed those events back and let bidding re-learn what a good lead looks like.
4. Make the form mirror your credit floor. Every field should map to a real underwriting criterion. A longer form that self-screens protects your officers’ time and raises the quality signal flowing back into your bidding.
5. Build the claim library with compliance as co-author, not gatekeeper. A component library with a fast lane turns a five-day review cycle into a same-day one, which is what makes testing — and therefore improvement — possible inside a regulated shop.

WeProms Digital has applied this compliance-aware PPC framework across Pakistani financial-services clients in SME financing, business banking, insurance, and fintech. The disclosures, qualification criteria, and review depth shift with each license and product — the sequence of audit, substantiate, realign, qualify, and feed back is what stays constant.