A high-ticket campaign once showed a return of roughly 30 times its ad spend, a result that would win more budget in almost any marketing review. The sales it took credit for came from people who were already planning to buy, which meant the campaign was capturing existing demand and generating very little of its own. Meta is now moving toward campaigns that run on little more than a product image and budget, with AI handling creative and targeting. That level of automation makes a captured sale even easier to mistake for a created one.
Siraj Dawar is Head of Marketing at Bee Enterprises, a company that has spent two decades launching international lifestyle brands in Pakistan. He has worked across 11 countries in South Asia and the GCC, running campaigns for high-ticket products where a single sale can move from social discovery to a showroom visit, phone negotiation and final search before it closes. On those campaigns, he skips the likes and impressions and follows the customer from the newsfeed all the way to the signed deal.
"The campaign that shows up on the dashboard is often different from the campaign that delivered," Dawar says. The 30x campaign was one he ran himself, and the dashboard's verdict didn't hold up once he moved budget into earlier-stage channels it couldn't track well. The real return surfaced in the telesales pipeline where his high-ticket deals closed, which is why he now judges every campaign by the demand it creates.
A 30x return can mean nobody new showed up
In a high-ticket purchase, the customer usually discovers the product in a social feed long before they ever reach checkout, and that first touchpoint rarely gets credit for the sale. "Attribution always rewards the last room the customer stood in, not the first room that convinced them," Dawar explains. Standard ROAS adds to the problem by counting sales that would have happened anyway alongside the ones a campaign influenced, and just over half of US brand and agency marketers now run incrementality tests and experiments to separate the two.
Inside a budget review, the pattern is easy to predict. The retargeting line with the gorgeous ROAS keeps its funding, while the prospecting line that fed that retargeting pool gets trimmed. Six months later, the pool is thinner and nobody in the room can explain why.
Automation scales whatever signal it's fed
Advantage+ and Performance Max learn from pixel and conversion signals, so a system fed mostly on existing customers gets very efficient at finding more people who look like them. "Automated bidding doesn't fix bad distribution. It scales it faster and gives you less visibility," Dawar says. "We don't fix the measurement problem with automated bidding. We just hand it to a faster system and ask it for numbers, whatever their quality."
When he joined Bee Enterprises, he inherited a campaign whose lead quality collapsed every time the budget went up. Tired creative explained part of the problem, but the audience was doing more damage. He pulled existing customers out of lookalike targeting and fed the CRM back into Meta as a list of people to avoid, a move that ran against the default performance marketing playbook.
"Tell the algorithm that you don't need these customers. You need new ones," he explains. "Anyone who has already interacted with our website, our social pages or our ads shouldn't be seeing that campaign."
Every campaign deserves its own scoreboard
Take a single campaign and measure it two ways. Run it for conversions, and Dawar follows the customer from a YouTube ad or newsfeed post to the closed sale, looking for the exact step where that path stalls. Run the same campaign for brand awareness, and the scoreboard shifts to impressions and how people come to think about the brand. The work before launch gets the same specificity. A pen sold to teenagers and the same pen sold to upper-income buyers need different selling points, and he matches each audience to the feature most likely to move it.
"A marketer who first understands the campaign, and what they want to learn from it next month, knows which metrics to analyze and optimize," Dawar adds. Across channels, he uses marketing mix modeling to compare performance side by side instead of letting each platform define success for itself. On a five-channel campaign, he pulls the metrics he cares about from each one, plots them together and looks for the channel where the slope starts breaking down.
In Pakistan and the GCC, the sale happens off the dashboard
Every framework above assumes the sale leaves a digital trace, and across much of the world it doesn't. In Pakistan, where Dawar is based, and throughout the GCC, a high-ticket purchase routinely moves from an ad into a WhatsApp chat and then onto a phone call, where trust gets built and the price gets negotiated. Meta has leaned into that behavior, and the revenue line driven by WhatsApp paid messaging topped $1 billion in a single quarter for the first time this summer, up 73% year over year.
"Most high-ticket selling here happens on WhatsApp and through telesales, because there's a trust issue to overcome and the financial conversation is a negotiation," Dawar notes. In the US and Europe, an ecommerce checkout often closes the loop inside the same system that served the ad. His teams track two journeys side by side, one showing how a person reacted to the creative and another that begins once they land in the conversation, and stitch them together by hand in the CRM.
Dawar still runs automated campaigns and still reads ROAS, treating both as a record of where a sale was captured. The dashboard can tell him where that capture happened, while the CRM and the WhatsApp threads show what created the demand behind it. . "Every day, you have to look at both journeys, find where they're breaking down and work out how to fix it," Dawar says.
The views and opinions expressed are those of Siraj Dawar and do not represent the official policy or position of any organization.
