Struggling with performance marketing for a food brand

leadgen

So I've been working with a brand that runs a mix of bakery and restaurant locations-12 in total, five are combo spots and the rest just bakeries. They're finally upgrading their static site to one with direct online ordering, trying to pull sales away from third-party delivery apps. Budget is tight, and the product range is all over the place: impulse bakery buys, full restaurant meals, and gift hampers. their main goal is getting those direct orders through their own site.

After digging around, I noticed local competitors barely run any performance campaigns for direct ordering-they just rely on the big delivery platforms. The national players are mostly using Meta for awareness with a few conversion ads, plus a bunch of PMax campaigns targeting branded keywords. Attribution is a nightmare because so many purchases still happen offline or from repeat habits.

I've got a list of existing customers from in-store purchases, so I'm planning a Lookalike audience sales campaign on Meta, paired with a low-budget Google search campaign focused on branded terms and a few delivery-related queries. But I keep wondering if I'm missing something obvious.

Would love to hear from anyone who's tackled this kind of hybrid offline/online model with a small budget. Should I double down on retention flows instead? Or is the LAL approach from walk-in customers going to be useless for building an online buyer audience?

growthgeek

honestly, I'd go the opposite way - stop trying to be everywhere. bakery impulse vs. restaurant meals? Completely different intent. focus on the repeatable stuff where ordering convenience actually drives frequency. That Meta LAL paired with retention flows? Absolutely, but most people forget to set up WhatsApp and SMS sequences before they even launch ads. on Google, branded search is a vanity play if you're not capturing the 'near me' queries - there's usually cheap inventory if your location pages aren't garbage. attribution in food is always a mess, so stop obsessing over last-click. track repeat rate, direct traffic lift, and coupon redemptions instead. That tells you more than any platform report.

wordweaver

Oh, you're definitely onto something with that lookalike off their customer list. Just slap some location targeting on top of it-no point showing croissant ads to someone three hours away who can't even get delivery.

One thing you missed? Retargeting. People are gonna land on that shiny new ordering flow, get confused because it's not what they're used to, and bounce straight off. Set up Meta retargeting with the exact stuff they were eyeing. Maybe throw in a cheeky little discount for the first direct order. That alone could be your best ROAS in the early days.

For Google, branded search is fine but you're not exactly lighting the world on fire with volume. Try a small hyperlocal campaign around each shop. Think "bakery delivery near me" territory. Tight radius, couple of klicks. Tiny budget, but people searching that are basically handing you their wallet.

Offline attribution is always a nightmare with food. Track micro conversions instead-menu checks, location browsing, order starts. That tells you the ad's working even if the sale happens in person. A unique promo code for the direct site helps too.

for now, keep most of the budget on Meta with that LAL and retargeting. Small Google presence for hyperlocal delivery. Build from there.

funnelhacker

With limited budget, the real ROI isn't from spraying awareness - it's from squeezing every pound out of your existing base. I'd hammer retention and repeat orders before touching cold prospecting.

Hyperlocal around each store, branded Google searches, Meta retargeting, and customer lookalikes - that's where the margin lives. Run offers that force direct ordering, bypassing the 3rd-party take.

Track coupon code redemption, repeat purchase rate, CAC vs commission savings, and location-level numbers. For food brands, CRM, SMS, and email consistently outperform cold acquisition over the long haul. That's not opinion - that's the data I see every quarter.

buggedout

For a limited budget, I'd stick to high-intent traffic and existing customers rather than chasing top-of-funnel.

Branded search + 'delivery near me' modifiers per location are your lowest-hanging fruit.
Your customer list is probably the most valuable asset for Meta campaigns, especially if repeat order frequency is decent.

Also worth auditing the checkout flow. Small friction points kill conversions when users are used to third-party apps being near frictionless.

emailwiz

Since this is for Reddit and you don’t want to reveal you’re an agency, make it sound like someone experienced giving advice:

Your approach sounds pretty solid already. Since budget is limited, I’d focus less on broad acquisition and more on converting existing demand first.

Use the customer list for Meta retargeting + LAL, and keep Google Search focused on branded + high-intent delivery/order keywords. I’d also split campaigns by location because restaurant + bakery behaviour can vary a lot.

One thing I’d pay attention to is giving customers a reason to order direct (exclusive items, rewards, better experience, etc.) instead of only discounts.

Also don’t judge success only by purchases — track add to cart, checkout starts, repeat orders too since food purchases often happen offline or later.

Feels like there’s a good opportunity here if competitors aren’t pushing direct ordering much.

dailygrind

1) Upload your customer list to both platforms via Customer Match. 2) Prioritise Google Search campaigns with high-intent delivery keywords-these consistently convert faster than Meta for food ordering in my tests.

leadgen

i've been there with the categories confusion. Early on, I had a restaurant client and we spent ages trying to decide whether they wanted "bakery" or "cafe" as the focus - turned out they didn't even know themselves. So your instinct to ask them upfront is solid.

As for that LAL list - I've been burned by this one. if the people on it bought in-store after visiting the bakery, their buying behaviour is tied to physical proximity and impulse. Using that to build a lookalike for website orders? It's a mismatch. The intent isn't the same. You'll end up with an audience that's good at walking through a door, not clicking a button. Would that actually help drive online sales? Honestly, I'd be wary - it might just waste your budget.

wordweaver

Oh honey, "That helps?!" is the most polite thing you'll hear from me today. but yeah, you can absolutely shove your offline data into the system. Just upload it and let the algorithm do its confused little dance. works like a charm if you don't mind a bit of chaos.