Lookalike Audiences for Event Lead Generation in Lahore

The Client

The client is an established Lahore event management company — an illustrative composite of the mid-size events businesses common across Punjab, not a named account. Founded in 2014, the company runs 38 full-time staff across design, production, and logistics, holds standing contracts with decorators, AV vendors, and caterers, and operates two storage yards in Lahore for staging, furniture, and décor inventory.

Its revenue mix is typical of the tier just below the national mega-agencies. Roughly 55% of revenue comes from corporate work — annual dinners, dealer conferences, product launches, award shows — about 30% from high-end weddings with a firm PKR 1.5 million minimum engagement value, and the remaining 15% from exhibitions and brand activations. Contracts run from PKR 1.5 million to PKR 12 million, averaging around PKR 3.2 million. Demand is intensely seasonal: corporate clients confirm 6-10 weeks ahead, with a Q4 budget-flush peak between September and December, while weddings cluster from November through April and book 3-6 months out. Summer is thin, which makes autumn pipeline volume an existential number, not a vanity one.

For a decade, growth came from the founding partners’ personal networks, a 46,000-follower Instagram profile that did the showcasing, and one marketing executive who handled the advertising: boosting the best-performing post each week at a budget of roughly PKR 600,000 a month. It worked, loosely, until it didn’t.

Two things forced the engagement. First, the calendar: the managing partner could see a quarter ahead that October’s pipeline was too thin to fill March’s wedding dates. Second, a blunt Q3 review found the sales team quoting fewer events than the prior year while ad costs rose, and nobody in the room could state what a lead actually cost. Businesses reaching that point usually discover what our event management marketing work surfaces first: the channel is rarely broken — the system around it is.

The Problem

Boosted posts had built the account’s habits, and those habits carried three costs the Q3 review finally quantified.

No measurement layer. The pixel had been installed in 2023 and fired exactly one event: PageView. There was no Lead event, no contact events, no way for Meta’s delivery to optimize toward anything commercial, and no retargeting pool deeper than casual video viewers. Ads Manager reported a cost per “result” that no one could map to a consultation, let alone a signed contract.

Leads everywhere, queue nowhere. Inquiries landed in page DMs, WhatsApp messages to three different personal numbers, and a Google Form embedded on a 2022-era landing page. Over a two-week sample, the team measured a 14-hour median first response. Event buyers in Pakistan contact two to four vendors the same evening; a 14-hour response is a forfeit, not a delay.

The junk ratio. Of roughly 200 raw inquiries a month, sales logged 38 as qualified — right event type, budget at or above the PKR 1.5 million minimum, date within 12 months. The other 81% were below-budget private parties, university fest organizers, job seekers, and vendors pitching decor services. At PKR 600,000 a month, the account looked like it was buying leads at PKR 3,000 each. Measured against qualified volume only, the real number was PKR 15,800 — a figure nobody had ever computed.

Paying to re-reach existing clients. With no exclusions anywhere, the boosted posts served heavily to followers — an audience dominated by past clients and their guests. Average frequency sat at 4.2 while new-reach share shrank every month.

The baseline, locked over the four weeks before the rebuild:

• Raw inquiries: ~200/month across DMs, WhatsApp, and the form
• Qualified leads: 38/month, a 19% qualification rate
• Effective cost per qualified lead: PKR 15,800 against a PKR 3.2 million average contract
• Consultations booked: ~23/month; attended: 8, a 35% show rate; signed: ~2
• Pixel events in use: PageView only; custom audiences in use: none

Phase 1 — Diagnosis, Tracking, and Seed Data Cleanup (Weeks 1-2)

Before touching audiences, the account needed to be able to see. The audit and its fixes, in dependency order:

Audit finding	Consequence	Fix applied
Pixel firing PageView only	No optimization signal, no lead retargeting	Instant-form Lead, WhatsApp click, and thank-you-page events deployed through GTM, with server-side Conversions API copies so browser restrictions could not thin the signal
One campaign mixing weddings and corporate	Creative averaged to nobody	Three segmented campaigns (Phase 2)
No exclusions on any ad set	Frequency 4.2, reach recycled to existing clients	Past-client and 180-day engager exclusion lists built
Intake across 3 personal inboxes, DMs, and a form	14-hour median response, leads lost	One WhatsApp Business number and instant forms piped into a single CRM board with pipeline stages
641 past events sitting in invoice history	The strongest audience asset, unused	Exported, normalized, deduplicated, and matched to Meta

Clean event instrumentation matters more in lead generation than most teams expect, because the events that define success happen off-platform. The form submission, the WhatsApp click, and the CRM outcome each needed a distinct, trustworthy signal before any audience strategy could be judged — the foundation described in our Meta Pixel and Conversions API setup scope, applied here to a service funnel rather than a checkout.

The seed data build was the quiet centerpiece. Three years of invoices yielded 641 past engagements with phone numbers and contract values. After normalization — deduplicating repeat corporate clients, standardizing +92 number formats — the file matched 590 records on Meta. Two seeds came out of it: all past clients, and a high-value segment of 187 engagements above PKR 3 million. Sales also signed off a written definition of “qualified” — event type in scope, budget band at or above PKR 1 million confirmed in the form, event date within 12 months — so that every later number meant the same thing to the ad account and the sales floor.

Phase 2 — Campaign Architecture and Audience Build (Weeks 3-5)

The rebuild replaced one boosted-post campaign with three, each owning a distinct buyer, and rebuilt intake so that what arrived was already sorted.

Layer	Before	After
Campaigns	1 boosted-post campaign	Corporate Events, Weddings, Exhibitions & Activations
Prospecting audiences	Followers + broad Lahore	1% lookalike of all past clients, 1% lookalike of high-value clients, interest-stack control set
Exclusions	None	All past clients + 180-day engagers on every prospecting set
Creative	2 boosted posts	6 rotating concepts, 2 per segment, refreshed weekly
Intake	DMs and personal WhatsApp	Instant forms with 3 qualifying questions + one WhatsApp Business number
Budget	PKR 600K undifferentiated	PKR 600K split 70/18/12 across corporate/weddings/activations

Three choices did most of the work that later got attributed simply to “lookalikes.”

The form asked the questions sales was asking. Each instant form carried three questions: event type, budget band as radio buttons from “under PKR 1 million” upward, and event month. Below-band submissions still arrived, but they arrived labeled, and sales stopped spending evenings discovering them. Raw cost per lead rose by design. The point of the form was to convert ad spend directly into qualified pipeline — the core discipline in our social media advertising engagements with service businesses, where the form is treated as part of the targeting.

The seeds were built on value, not just identity. A 1% lookalike of 590 past clients models what the average client looks like. A 1% lookalike of the 187 engagements above PKR 3 million models what the clients worth having look like. The two audiences overlapped far less than expected and, later, behaved very differently on cost and close rate.

Creative spoke to committees and families, not “event planners.” Corporate prospects buy production reliability: a 40-second recap reel of a 1,200-guest dealer conference, a founder talking-head on contingency planning and backup power. Wedding prospects buy emotion: a cinematic recut of a single wedding, 30 seconds, sound on. Same brand, two vocabularies, zero cross-showing — and each creative ended with a segment-specific form rather than a generic “learn more.”

Phase 3 — Optimize, Prune, and Scale Winners (Weeks 4-8)

With three campaigns and four audience classes live, optimization became an economics exercise. Two rules ran from week 4: any ad set exceeding twice the target qualified cost per lead for three consecutive days with 8,000+ impressions was paused, and any creative that beat the campaign median for two weeks spawned a variant pair for testing. The creative team produced two new concepts per segment per week — volume matters because event creative fatigues in 10-14 days in this category.

Cost per qualified lead by audience, week 4 versus week 8:

Audience	Week 4	Week 8	Budget share, week 8
Interest-stack control	PKR 15,900	PKR 14,100	15%
1% lookalike, all past clients	PKR 11,900	PKR 8,900	40%
1% lookalike, high-value clients	PKR 10,200	PKR 7,600	25%
Retargeting (form openers, 75% video viewers)	PKR 6,400	PKR 5,300	12%

The control stack kept its budget deliberately. It was the honest benchmark — without it, the lookalike gains would have been a story rather than a measurement. By week 8 the high-value lookalike produced qualified leads 46% below the control’s cost, but at low volume, which is the standard trade. Budget migrated to the 1% all-client lookalike as the volume carrier, with the high-value set and retargeting as efficiency ballast. The remaining 8% of budget funded a Karachi pocket test for destination-wedding demand — promising enough to keep, small enough to protect the account’s Lahore learning phase.

Two operational changes on the sales side landed inside the same window. First, a response SLA: every new inquiry answered within one hour during business hours, enforced by the CRM board, with after-hours inquiries answered first thing the next morning. Median first response fell from 14 hours to 47 minutes by week 6. Second, a WhatsApp confirmation flow the day before each consultation — location pin, time, the consultant’s name. Show rate climbed from 35% to 52% over six weeks. Neither change touched the ad account; both changed what the ad account’s money bought.

Phase 4 — Offline Feedback and Seasonal Compounding (Weeks 8-12)

The last lever was closing the loop between Meta and the CRM. From week 8, every week’s CRM export was uploaded as offline conversion events — Qualified, Consultation Booked, and Signed, with contract values attached. By then the account had accumulated 75+ imported booked-consultation events, enough signal for delivery to optimize toward consultation bookings rather than form submissions. That is when cost per qualified lead fell through the PKR 10,000 floor and stayed there: PKR 9,760 for the final 28-day window.

The lookalike ladder extended at the same time. The 2% and 3% tiers of the all-client seed added reach at a cost roughly 9% above the 1% tier — an acceptable premium for the +40% volume that let the wedding campaign carry a real booking season rather than a trickle. Exclusions grew with the list: every newly signed client left the prospecting audiences within a week, which kept frequency honest as spend rose.

In week 10, the seasonal flight the whole quarter had been building toward opened: a wedding-season campaign for November-February dates, running the two proven creatives against the proven audiences, aimed at the 3-6 month booking curve rather than last-minute demand. Spend scaled from PKR 600,000 to PKR 800,000 a month across weeks 6-10 — not as a growth gesture, but because the qualified-cost floor held while volume scaled, which is the only sane reason to raise a lead-gen budget.

The final 28 days against the pre-engagement baseline told the whole story: spend up 33%, raw inquiries flat, qualified leads up 116% — and sales signing 7-8 events a month from paid against roughly 2 before, at a PKR 3.2 million average contract value.

Final Results

Measured at day 90, trailing 28-day window against the four-week pre-engagement baseline:

Metric	Baseline	Day 90	Change
Monthly ad spend	PKR 600,000	PKR 800,000	+33%
Raw inquiries per month	~200	~205	flat
Qualified leads per month	38	82	+116%
Qualification rate	19%	40%	+21 pts
Cost per raw lead	PKR 3,000	PKR 3,900	+30%
Cost per qualified lead	PKR 15,800	PKR 9,760	-38%
Consultations attended per month	8	26	+225%
Signed events per month from paid	2	7-8	~3.5x
Cost per signed event	~PKR 300,000	~PKR 108,000	-64%

Every row traces to a phase: the qualification rate to the Phase 2 form, cost per qualified lead to the Phase 3 audience economics, the consultation row to the response SLA and confirmation flow, and the final cost floor to the Phase 4 offline imports. One honest note: raw cost per lead rose 30%, and if your dashboard stops at raw CPL, this rebuild would have looked like a failure in week 3. These figures describe an illustrative composite engagement — representative of the outcome shape this framework produces for established Pakistani service businesses, not an audited client account. Sanity-check them against your own seed-list size, average contract value, and sales response capacity before extrapolating.

What Made This Work

1. The invoice file was the real media asset. 641 past engagements with contract values let Meta model the actual customer, and the 187-record high-value seed modeled the customer worth acquiring. No interest stack in Ads Manager approximates either, and the control set proved it weekly.

2. Qualification moved upstream into the form. Filtering at the form costs question design; filtering on the sales floor costs consultant hours and morale. Raw cost per lead rising 30% while effective cost per lead fell 38% is the honest arithmetic of that trade.

3. Offline conversions closed the optimization loop. Until booked consultations were imported weekly, the ad account was optimizing toward people who fill forms. Afterward it optimized toward people who sign contracts — the single biggest week-over-week cost drop followed that change, not any audience edit.

4. Segment creative respected the two buyers. A procurement committee vetting a dealer conference and a family booking a wedding are different purchases with different anxieties. Sharing one vocabulary across both was the original account’s quietest and most expensive mistake.

5. Speed was treated as a paid-media variable. A 47-minute first response against a 14-hour one changes conversion on identical lead flow. The ad account did not produce dramatically more raw leads in month three — the sales floor simply stopped forfeiting the ones it bought.

What Teams Can Apply

1. Mine your invoice history before your interest targeting. If you have 500+ past customers with contact details and order values, your strongest audience work is a data-cleaning exercise, not a targeting-tools exercise. Segment the seed by value before building any lookalike, and keep a control set so the gains are measured, not asserted.

2. Report cost per qualified lead, or report nothing. Raw cost per lead flattered this account for a year while it paid PKR 15,800 per real opportunity. Define “qualified” in writing with sales, then hold the ad account to that number in every weekly review.

3. Expect raw cost per lead to rise when you add qualifying questions. That is the mechanism working — fewer total submissions, more real ones. Judge the change over 4-6 weeks of qualified volume, not the first week’s cost column.

4. Import offline outcomes weekly once your CRM tracks them. It is the cheapest optimization upgrade available to any service-business lead account, and it compounds with audience work rather than replacing it. Most Pakistani service businesses already have the data; they just never connect it to the ad platform.

5. Buy against your booking curve, not the calendar. Event demand books 6 weeks to 6 months ahead of delivery. The campaigns that filled March were opened in October — spend scheduled against the decision window, not the delivery window, is what turns seasonality from a cash-flow crisis into a schedule.