By Hamza Ali · WeProms Digital · November 10, 2026
Last updated: November 2026.
A halal caterer in Dallas spends $2,300 a month on Instagram and Facebook ads for her shaadi and corporate catering business (worked example — substitute your own numbers). Meta Ads Manager reports 30 leads that month at roughly $45 per lead. The bank account tells a different story. The two biggest deposits came from a wedding booked through a mosque WhatsApp group and a corporate iftar that a former client forwarded to her office manager on Zelle deposit day.
When we audit marketing for Pakistani-American food businesses in Dallas, Plano, and Irving, this gap is the first thing we check. Here’s the thing. The ad platform is not lying, exactly. It counts every lead it touched — including customers who were already coming and simply clicked an ad on the way.
The dashboard says one thing. The community says another.
The channels that actually fill a desi catering calendar are invisible to every analytics tool you own. Nobody signs a wedding contract because a retargeting pixel followed them. They sign because an auntie in the family WhatsApp group sent photos from a valima, because a mehndi artist keeps a shortlist of three caterers, or because someone mentioned your biryani outside Friday prayers at the masjid. Researchers call this dark social — sharing that happens in private channels like WhatsApp, texts, and voice notes, which analytics tools cannot see or credit. If your customer’s journey started in a group chat, Google Analytics records it as “direct” or hands the credit to whatever ran last.
Buyers do not travel one road anymore, and the data backs that up. A Semrush survey of 1,030 U.S. consumers who had tried AI tools found 77% use AI tools alongside search engines when researching online, moving between search, social, video, and AI answers before they ever call a business. So what? Your future catering client may meet your name in four places before the first phone call, and the platform that gets the last click rarely deserves the whole booking. One caveat: that survey describes American consumers broadly, not halal catering demand in Dallas — it proves journeys are mixed, not that your mix matches it.
Ad dashboards systematically tilt credit toward the channels they can measure. MarTech, summarizing how enterprise teams now combine media mix modeling with post-purchase surveys, puts it plainly:
“Click-based tools tend to give too much credit to bottom-of-funnel channels like branded search or retargeting because those are often the last measurable touchpoints before conversion.”
That is the mechanic. The ad that catches a warm lead at the end gets the receipt, and the WhatsApp group that did the actual selling gets nothing. Judging your marketing by the ad dashboard alone is like judging your Eid weekend by the till on Monday morning: the money is real, but the story of where it came from lives in a hundred private thank-you messages you will never see.
Where the $2,300 actually goes
Run the unit economics before you argue about attribution. Our Dallas caterer’s typical family event bills 120 guests at $65 per head, or $7,800. Here is the worked example event P&L, labeled illustrative:
| Line | Worked example (120 guests at $65/head) |
|---|---|
| Event revenue | $7,800 |
| Ingredients and zabiha meat | $2,960 |
| Chefs and staff | $1,540 |
| Transport, rentals, disposables | $460 |
| Contribution per event | $2,840 |
Contribution — what remains after the direct costs of delivering the event, before marketing and overhead. Every marketing argument should be settled against that $2,840, not against clicks.
Now apply it. Meta claims 30 leads at $45 from the $1,350 lead-campaign portion of the budget. If the survey later shows only 9 of those 30 bookings genuinely started with the ad, the real cost per ad-sourced booking is $150 — still comfortably profitable at $2,840 contribution, but triple what the dashboard implies. The remaining $950 of the budget ran retargeting and boosted posts, mostly re-touching people who already knew her. The fix is different for each half: measure the first, cut most of the second.
The one-question survey that settles it
You do not need new software to close the measurement gap. A post-purchase survey — a short question asked after a customer has committed, not while they are still deciding — is enough. Add one required field to your booking form: “How did you first hear about us?” Offer fixed options plus a blank line. MarTech’s guidance on survey design applies directly here: mix structured options with open-text responses, so customers can name the specific WhatsApp group, the specific auntie’s wedding, or the specific flyer — that detail is where the real answers live.
Ready to paste into your form or WhatsApp Business quick reply:
“How did you first hear about us? (a) Someone told me (b) A WhatsApp or community group (c) Instagram or Facebook ad (d) Google search (e) Walked past, event, or flyer (f) Ordered from us before — or tell us in your own words.”
In the worked example, the caterer captured 86 responses over 60 days — every confirmed booking got asked once, by form or in the confirmation message, and the tally took one spreadsheet column. Expect this to be boring. That is the point. The boring column beats the exciting dashboard because it records what customers say, not what a platform claims.
What sixty days of answers changed
The survey split looked like this: 38% word of mouth through family, friends, and the mosque; 22% WhatsApp and community groups; 14% Google search; 12% Instagram and Facebook ads; 9% walk-past, events, and flyers; 5% repeat clients. So what? More than half the calendar came through channels no dashboard can name, and the ads’ honest share was 12% — not the 100% the lead report implied.
Three moves followed. First, she kept the ads but rewrote them as reminders for people who already knew the name, because that is the job they were actually doing. Second, $500 of the retargeting budget moved to sponsoring community iftars and masjid fundraisers, where her actual customers already gather — spend aligned with the real top of the funnel. Third, before the winter shaadi season and the Ramadan corporate-iftar rush (Ramadan is expected to begin around 8 February 2027, subject to moon sighting, with corporate enquiries starting about six weeks earlier), she stopped counting leads and started counting booked events per source. Based on our experience reviewing accounts like this, most community food businesses find ads directly started fewer than half their bookings — if your survey disagrees, your market or your ads genuinely differ, and that is worth knowing too.
One warning if you also run Google Ads. Reporting published November 9 found that batch uploads of offline conversions across multiple accounts with identically named conversion actions can credit conversions to accounts that never generated the click — a “successful” upload only means the file processed. Upload conversions one account at a time, use unique conversion names, and reconcile against click IDs. The same skepticism you apply to Meta’s numbers applies to every platform’s self-report.
The 15-minute fix
- Add “How did you first hear about us?” as a required field on the booking form, with the six options above plus an open line.
- Save the same question as a WhatsApp Business quick reply for clients who book by chat, and note the answer before you send the Stripe or Square deposit invoice.
- Log every confirmed booking’s source in one spreadsheet column — a free Google Form feeding one sheet is enough.
- At day 30 and day 60, pull the ad platform’s claimed conversions for the same window and put the two numbers side by side.
- Recalculate cost per booked event per source, against your own contribution math, not against clicks.
- Reallocate 20% of the ad budget toward the top two real sources before the next shaadi season, and keep the survey running.
The decision rule is simple once the split exists. If the survey and the dashboard agree, scale the ads with confidence. If they disagree by more than half, fix the measurement before adding a single dollar to spend — otherwise you are funding a story, not a channel.
Read next: Why Meta reports sales your meal-prep store never made and how to track Google Ads leads all the way to shaadi bookings.
We have worked with halal caterers, restaurants, and grocery businesses across the US, UK, and Gulf on exactly this problem — turning one spreadsheet column into a real measurement system that decides budgets. If you want the marketing measurement and reconciliation setup we run for overseas Pakistani businesses, or call tracking that connects phone bookings to the ads that earned them, WeProms Digital will build it in your market currency and your timezone. Message hello@weproms.com or WhatsApp +92 300 0133399, or use the contact page — replies come in US business hours.
Frequently Asked Questions
How much does a “where did you hear about us” measurement setup cost with WeProms?
The do-it-yourself version costs nothing but one spreadsheet column. A WeProms build — survey design, booking-form integration, source tagging, and a dashboard that reconciles ad-platform claims against survey truth — is scoped per business after a short call, in USD for US clients. We never quote package prices in articles; the contact team prices it for your event volume.
Is asking customers where they heard about us allowed under US privacy rules?
Yes. A source question with no personal details attached is ordinary first-party information you collect from your own customers. Keep answers anonymous in your tally, do not attach them to payment data you do not need, and honor any customer who skips it. No special consent machinery is required for an optional one-line question.
How many responses do I need before trusting the split?
Around 30 booked events gives you a reliable direction; 60 days of bookings is better. Treat the percentages as ranges, not verdicts — measurement writers at MarTech make the same point about unstable marketing processes generally: ranges reflect how the work actually behaves, single precise numbers overstate certainty.
Why does Meta claim more customers than my survey shows?
Meta counts view-through and last-click credit, so anyone who saw an ad and later booked through a WhatsApp group can appear as an ad lead. Your survey asks the customer directly, which is a different, more honest question. The two numbers answer different things; budget against the survey.
Should I stop Instagram ads if the survey shows mostly word of mouth?
Usually no. Ads that remind a community which already knows you still earn their keep — at reminder pricing, not lead-hunting pricing. Restructure the campaigns around the reminder job, cut the retargeting layer that chases existing fans, and put the savings where the survey says demand actually starts.
About WeProms Digital
WeProms Digital is a Lahore-based marketing measurement and operations agency working with overseas Pakistani and halal-minded business owners across Dallas, Houston, Toronto, London, Dubai, and beyond.
The team specialises in marketing measurement, call and offline conversion tracking, and analytics setups that survive scrutiny from real bank statements, with a track record of reconciling platform-reported leads against what customers actually say.
Get in touch: hello@weproms.com · WhatsApp +92 300 0133399 (replies in US, UK and Gulf business hours) · weproms.com/contact-us
Sources & References
- MarTech — Combining MMM with post-purchase survey attribution — November 9, 2026
- Semrush Blog — The future of SEO: Where search is heading in 2027 — November 9, 2026
- MarTech — Why marketing needs different approaches to measurement — November 9, 2026
- Ignite Visibility — Franchise and Multi-Location Marketing News — November 9, 2026
- WhatsApp Business — business messaging platform — accessed November 2026
- Stripe — pricing — accessed November 2026
- Google Forms — free survey tool — accessed November 2026
Additional reading from industry feeds:
