Real Estate Google Ads Case Study in Pakistan

The Client

A Lahore-based residential property developer selling plotted land, constructed villas, and apartment units in a mid-scale gated community on the city’s expanding outskirts. The project targeted middle and upper-middle income buyers — families looking to build on 5-marla and 10-marla plots, and professionals buying one- and two-bedroom apartments as either a first home or an investment asset. Typical ticket sizes ranged from roughly PKR 8.5M for a 5-marla plot up to PKR 32M for a constructed villa, with apartment units sitting between PKR 11M and PKR 19M depending on floor and view.

The developer had a competent in-house operation: fourteen relationship managers, a working CRM, a site office running daily visits, and a modest performance marketing function run by a single in-house marketer. Over the previous year, that marketer had steadily scaled Google Ads spend, operating on the reasonable assumption that more budget meant more buyers. Lead volume had indeed climbed. The sales team’s complaint, however, was constant and specific — the leads were unbuyable. Brokers fishing for commission splits. Buyers with a PKR 5M budget enquiring about PKR 30M villas. Enquirers based in other cities who would never travel for a site visit. Diaspora investors who expected a different conversation entirely.

Out of roughly 320 form fills arriving each month, the team counted fewer than 90 qualified site visits. The developer was spending PKR 2.4M a month on Google Ads and could not say with any confidence whether that spend was producing revenue or simply generating activity. When WeProms Digital was engaged, the brief was deliberately constrained: hold the media budget flat, but make the leads buyable. This case reviews the 90-day restructure that followed.

The economics made the problem urgent. Each relationship manager could realistically run eight to ten live site-visit cycles a month — beyond that, follow-up suffered and conversion dropped. With fourteen managers, the team’s effective monthly capacity sat around 120 to 140 productive site visits. The campaign was already generating more raw leads than the team could work well, and the surplus was almost entirely unqualified noise burning through their attention. The lever was not more leads; it was a higher proportion of leads worth calling. A 34% reduction in cost per qualified lead at flat spend would, on paper, fund roughly forty additional qualified enquiries a month — still inside the team’s capacity — without a single additional rupee of media. Reframing the goal around qualified-lead economics, rather than lead volume, set the terms for everything that followed.

The Problem

Four issues were destroying lead quality before any sale could happen.

1. Keyword targeting built for reach, not intent. The account ran on broad-match keywords like “plots in Lahore,” “property for sale,” and “houses in Pakistan.” These terms capture enormous search volume, but most of that volume is informational, broker-driven, or geographically irrelevant. The developer was paying for clicks from people researching market trends, not from people ready to book a plot.

2. A single landing page for every ad group. Fourteen distinct ad groups — covering plots, villas, apartments, and investment packages — all sent traffic to one generic, homepage-style landing page. A buyer searching “10-marla plot price” landed on a page describing the whole community with no plot-specific price, no availability map, and no payment plan. Bounce rates sat above 68%, and the page converted the same regardless of intent.

3. No conversion feedback loop. Google Ads was optimizing toward form submissions. But a form submission in real estate is barely a signal — it is the start of a conversation, not a sale. With no data on which leads actually booked a site visit, placed a refundable token, or closed, the bidding algorithm kept doubling down on the cheapest leads to submit forms: low-intent, low-budget, and broker traffic.

4. Slow lead response killed hot leads. Inbound leads were emailed to a shared sales inbox and distributed manually each morning. Median first-response time was nine hours. In Pakistani real estate, a buyer who enquires in the evening has usually contacted two or three competing developers by breakfast. Hot leads were going cold before the first call.

Phase 1 — Account Audit and Tracking Repair (Weeks 1–2)

The first two weeks were spent understanding what the account was actually doing and fixing the measurement gaps that made optimization impossible.

Spend and keyword audit. We pulled twelve months of search-term data and reclassified every click against the developer’s actual unit types and price tiers. The picture was stark:

Keyword theme	Spend share	Qualified-lead share
Generic “plots / houses Lahore” broad match	54%	22%
Unit-specific (“10 marla plot price”)	18%	41%
Brand / project name	12%	19%
Competitor / broker terms	9%	4%
Diaspora investor (“property in Pakistan from UAE”)	7%	14%

The generic broad-match bucket was consuming more than half the budget while producing barely a fifth of qualified leads. Conversely, the unit-specific and diaspora buckets were starved — underfunded segments producing disproportionate value.

Negative keyword build-out. Before restructuring anything, we added 480 negative keywords: broker terms (“dealer,” “agent,” “commission,” “referral”), job and price-research terms (“jobs,” “salary,” “trend,” “forecast”), and out-of-budget modifiers (“cheap,” “low price,” “installment under 1 lakh”). We also excluded a set of adjacent-intent terms the developer had been paying for — rental searches, commercial-office enquiries, and plot-size conversions. This cleanup alone cut wasted spend by an estimated 19% within the first ten days, freeing budget for the segments that actually produced site visits.

Offline conversion tracking — the foundation. This was the single most important fix in the engagement. We integrated the sales CRM with Google Ads so that every lead could be scored and pushed back upstream as an offline conversion, in three tiers:

• Tier 1 (Site Visit Booked): lead booked a physical site visit — the strongest pre-sale signal in real estate.
• Tier 2 (Qualified Enquiry): the relationship manager confirmed correct budget, city, and purchase timeline on the first call.
• Tier 3 (Token / Booking): a refundable token was placed — a hard revenue signal.

The broker problem deserved special attention because it is peculiar to Pakistani real estate. A meaningful slice of property search traffic comes from independent brokers scouting inventory to resell, not from end buyers. Broker traffic inflates click volume and form submissions — both of which the old bidding strategy rewarded — while producing almost no site visits and no tokens. We built a broker-detection layer using recurring phone numbers, broker-pattern keyword combinations, and CRM disposition flags, and treated confirmed-broker leads as a negative signal rather than a conversion. Over the first month this removed a stubborn class of low-quality traffic that manual filtering had never quite eliminated.

These tiers were uploaded daily via the Google Ads API and configured as separate conversion actions with different values. From this point forward, Smart Bidding optimized toward site-visit bookings and tokens, not form fills. Everything in the subsequent phases depended on this loop being clean.

Phase 2 — Restructure and Segmentation (Weeks 3–5)

With measurement repaired, we rebuilt the account around the buyer’s actual intent and price tier.

Campaign architecture. We collapsed the bloated structure into three clean campaigns, each with tightly themed ad groups. A plots campaign held 5-marla, 10-marla, and 1-kanal ad groups, each with dedicated match types and budgets. A constructed-units campaign separated villas, apartments, and investment packages by price tier. A diaspora and investor campaign targeted overseas buyers by location (Gulf, UK, US) and interest audiences, with messaging tuned to currency-neutral pricing and overseas documentation.

Audience overlap — the silent efficiency killer where multiple ad sets compete for the same user — was eliminated using shared negative audience lists so a given searcher could only enter one funnel. This mattered more than it sounds: before the restructure, a returning villa prospect was simultaneously eligible for six ad groups, each bidding against the others and inflating effective CPCs.

Match-type and bid strategy. Broad match was replaced with a phrase-and-exact foundation, with broad match retained only where Smart Bidding had accumulated enough Tier 1 and Tier 2 conversion data to constrain it responsibly. Bidding moved from manual cost-per-click — which had been optimizing for clicks — to Maximize Conversions targeting the Site Visit Booked conversion, with a target CPA layered in once two weeks of clean data had accumulated.

Mid-point results at week 5:

Metric	Pre-engagement	Week 5
Cost per qualified lead (PKR)	9,800	7,900
Unqualified lead share	71%	48%
Site-visit booking rate (paid leads)	14%	19%
Wasted spend on broker / out-of-budget	~31%	~9%

The account was already moving in the right direction, but lead quality was still being capped by the landing-page experience — which is where Phase 3 began.

Phase 3 — Landing Pages and Creative (Weeks 4–7)

Better targeting only works if the destination page converts. The single generic landing page was replaced with a unit-specific landing-page system.

One page per intent. We built eight dedicated landing pages, each matched to a specific ad group and price tier. The 5-marla plot page carried the plot price band, an interactive availability map, and the payment plan. The 10-marla page added corner-premium pricing and current possession dates. Villa pages carried floor plans, a possession timeline, and a gallery of constructed units. Apartment pages showed floor plans, view direction, and an indicative rental yield for investor traffic. A dedicated diaspora page presented currency-neutral pricing and the overseas-buyer documentation process front and center.

Each page carried the specific price band — a deliberate choice, since Pakistani property buyers who see a price self-qualify far more reliably than those who are told to “enquire for price.” Each page also carried a form shortened from eleven fields to four (name, phone, unit interest, budget band), and each form’s submission fed the right CRM pipeline directly.

Creative refresh. The generic “luxury living” image ads were replaced with unit-specific creative: real plot maps, actual floor plans, and price-led headlines (“10-Marla Plots from PKR 8.5M — Corner Plots Available”). Ad copy leaned on the signals Pakistani property buyers actually weigh at decision time — on-ground development progress, possession dates, and documentation confidence. Reducing the form from eleven fields to four alone lifted submission rate 27%, and because those submissions were now tied to a specific unit intent, the leads were substantially more qualified.

Site-visit booking as the primary CTA. The primary call-to-action shifted from “Enquire Now” to “Book a Free Site Visit,” backed by an immediate calendar slot picker. This raised the intent bar at the point of conversion — the people who submitted the form were already committing to a visit, which is precisely the Tier 1 signal the bidding algorithm now valued.

Phase 4 — Offline Conversion Optimization and Scale (Weeks 6–12)

With the funnel rebuilt and tracking live, the final phase let Smart Bidding do its job — and then scaled what worked.

Value-based bidding. Once each conversion tier had accumulated roughly 50 conversions per month, we switched the plots and villa campaigns to Maximize Conversion Value, weighted by the token-booking value. The algorithm began favoring the lead segments that actually produced revenue rather than the cheapest form fills. The diaspora campaign, where conversion volume was lower but ticket value higher, stayed on target-CPA toward site-visit bookings to avoid being starved by the value model.

Modeling the token value required care, because a villa token and an apartment token are not the same event. We assigned each tier a blended expected value derived from historical close rates and average ticket — a villa site visit was weighted substantially higher than a 5-marla plot enquiry — so the value model would not cheaply chase low-ticket volume at the expense of the unit types that actually drove revenue. We also layered audience expansion cautiously: once the core intent audiences were saturated, we opened similar-segment expansion only for the villa and 10-marla campaigns, where margin could absorb the slightly higher acquisition cost, and left the lower-tier campaigns on tight intent targeting. This kept the account efficient rather than letting the algorithm trade quality for volume the moment it had bidding freedom.

Lead response automation. In parallel with bidding, we wired the form into an instant-notification flow: every submission triggered an SMS and WhatsApp message to the assigned relationship manager within 60 seconds, with the lead’s stated budget and unit interest pre-filled into the CRM. Median first-response time fell from nine hours to 14 minutes. In real estate, where the first responder usually wins the meeting, this measurably lifted the site-visit booking rate on its own — paid leads were now being worked while still warm.

Scale test. In weeks 10–12 we increased spend on the two strongest ad groups — 10-marla plots and villas — by 35%. Because the conversion signals were clean, efficiency held: cost per qualified lead rose only 6% while qualified-lead volume grew 31%. This is the test that matters in lead generation: not whether you can spend more, but whether efficiency survives the spend.

Final Results

Metric	Before	After (90 days)	Change
Cost per qualified lead (PKR)	9,800	6,470	-34%
Qualified enquiries per month	88	139	+58%
Lead-to-site-visit rate	14%	26%	+86%
Unqualified lead share	71%	33%	-54%
Median first-response time	9 hours	14 minutes	-97%
Site-visit bookings from paid	90	192	+113%
Offline conversions fed to Google Ads	0%	100%	—

All figures are illustrative composites built from the patterns WeProms sees in Pakistani real estate campaigns, intended to help a buyer sanity-check fit — not audited third-party results.

What Made This Work

1. We changed what the algorithm was optimizing toward. The turning point was offline conversion tracking. Until Google Ads could see which leads booked site visits and placed tokens, every bid strategy was flying blind toward cheap form fills. Feeding Tier 1, 2, and 3 signals back into the account re-routed spend toward buyers, not enquirers. Every later gain compounded on this foundation.

2. Negative keywords did the heavy lifting before any restructuring. Cutting 19% of wasted spend on broker and out-of-budget traffic in week one freed budget for the unit-specific terms that actually produced qualified leads. Restructuring an account full of junk traffic would only have produced a cleaner junk account.

3. Landing-page-to-ad-group match fixed the leak in the middle of the funnel. Better targeting just sends the right person to the wrong page faster. One unit-specific page per intent — with real prices, an availability map, and a four-field form — turned qualified clicks into qualified leads.

4. Speed-to-lead is a bidding problem in disguise. Nine-hour response times meant the campaign’s best leads were spoiling before sales ever touched them. Getting response under 15 minutes raised the effective value of every conversion already in the system, without spending another rupee on media.

What Teams Can Apply

For Pakistani property developers running paid lead generation:

1. Score every lead and feed it back to Google Ads as an offline conversion. If your bidding strategy optimizes toward form fills, it will find you the cheapest form-fillers in the country — not buyers. Robust offline conversion tracking is the foundation everything else stands on.

2. Run negative-keyword discipline before you restructure. Pakistani real estate search terms are saturated with broker, job, and price-research intent. A clean 400-plus negative list will quietly fund your best ad groups.

3. Match one landing page to one buyer intent, with the actual price on it. Pakistani property buyers who see a price are far more qualified than those who see “enquire for price.”

4. Treat speed-to-lead as part of the media strategy, not a sales-team afterthought. A lead that goes cold in nine hours wastes the click you paid for. Automate first-touch under 15 minutes.

5. Right-size lead volume against sales capacity before raising budgets. More leads than your team can call well is wasted spend. Calculate productive site-visit capacity per manager, then engineer the campaign to fill it with qualified enquiries rather than overwhelm it with raw volume.

WeProms Digital has applied this Google Ads management framework across Pakistani property developers in Lahore, Islamabad, and Karachi. The unit types, price tiers, and geographic segments change with each project — but the principle holds: optimize toward the conversion that predicts revenue, not the one that is easiest to count. For developers exploring the broader playbook, the real estate marketing hub covers the full funnel from awareness to booking.