How should vegan restaurants measure Meta and Instagram ads ROI when most covers are walk-in or first-party reservations?

Vs other guides

Guide	Job
Paid social for sustainable brands	How to run product ads and proof creative. Not dining-room measurement.
Incrementality testing	Geo and SKU holdouts for DTC. A 12-table restaurant cannot copy that design.
iOS attribution	Shopify Purchase via CAPI. Restaurants need offline events, not checkout Purchase.
ROAS benchmarks	DTC ROAS floors. Dining uses cost-per-cover versus contribution.
Weekday covers	School-night bundle math. Not paid social.
Vegan restaurant Google listing	Organic Maps. Do not mix GBP posts into this ROI sheet.
Vegan hotel direct bookings	Direct vs OTA mix. Hotels reuse this worksheet with ADR.
This guide	Contribution floor, cost-per-cover cap, weekly offline upload, then the social ROI calculator

Why platform ROAS cannot see covers

Meta and Instagram credit a click, a landing-page view, a reservation thank-you if you installed one, or a message thread. They do not see a walk-in who saved a Reel and sat down on Saturday. They also miss first-party reservations that finish inside OpenTable, Resy, Tock, SevenRooms, or a branded widget that never fires your pixel.

Two errors stack. Ads Manager can over-credit diners who would have booked from Maps. It can also under-credit walk-ins the pixel never meets. Neither error is a 3x ROAS. Boosts count as spend. Instagram DMs are not covers until a reservation or POS check exists. Optimize on booked covers you can upload, not profile visits.

Steal this worksheet

Numbers below are a worked example. Replace every cell from POS and Ads Manager before you change budget. Do not invent occupancy lift or a 3x target.

Line	Formula	Worked example	Your cell
Average ticket	POS, last 4 weeks, dine-in + first-party pickup	$28	[avg_ticket]
Food cost %	House recipe or P&L (many independents use 28 to 32%)	30%	[food_cost_pct]
Contribution per cover	ticket × (1 − food %)	$19.60	[contrib]
Weekly Meta + Instagram spend	Ads Manager, ads + boosts	$400	[weekly_spend]
Contribution floor	spend ÷ contribution = break-even covers	21 covers	[floor]
Cost-per-cover cap	Leave room for labor and overhead. Example: contribution minus $8	$11.60	[cap]
Covers needed at cap	spend ÷ cap	35 covers	[needed]
Matched covers this week	Reservations and POS checks you can tie to an email or phone in the upload	Fill from CRM	[matched]
All covers this week	POS, including walk-ins you cannot match	Fill from POS	[all_covers]
Attributed revenue	matched covers × ticket	Fill	[attributed_revenue]

Include pickup packaging in food cost. Skip delivery-app commission in this sheet. Marketplace fees are a second tax and belong on a separate mix decision. Time labor in the break-even calculator if you staff extra hosts for the ads.

Two gates, not a 3x rule

Gate	Math	If it fails
Optimistic cost per cover	spend ÷ all covers	Kill or pause. Even if every diner came from ads, contribution does not cover spend.
Matched cost per cover	spend ÷ matched covers	If this sits above the cap but the optimistic gate still passes, fix matching for two weeks before you scale. Do not assign unmatched walk-ins to Meta.

The floor is the food-contribution break-even (21 covers in the example). The cap is the number you will actually pay (35 covers in the example). If matched covers beat the floor but miss the cap, you funded ads with labor and rent. Cut spend. Do not raise budget because Ads Manager shows a high ROAS.

Host-stand "Instagram" tallies are a triangulation note. They are not an offline event and they are not attributed revenue.

Weekly offline conversion upload

Create an offline event set in Meta Events Manager on the ad account that spends. Upload Monday for the previous week. Daily is better. Independents who wait a month teach last month's menu.

• event_name: Purchase
• event_time: reservation or POS close
• value and currency: check total
• email and/or phone: from the booking tool or POS CRM
• order_id or reservation id: so a re-upload does not double-count

Upload first-party reservations with an email, and POS checks already in your CRM. Do not invent walk-in emails or paste "how did you hear" ticks. Incomplete matching is honest. Fake rows poison optimization. Fire pixel Purchase on the reservation confirmation URL you own. If the booking tool already sends Purchase with email through Conversion API, skip a duplicate CSV for those IDs.

Then open the social media ROI calculator

The social media ROI calculator is the shareable arithmetic after the worksheet, not instead of it. Platform: Meta (or Instagram if you split). Spend: [weekly_spend]. Revenue: [attributed_revenue] from POS-matched covers, never Ads Manager. Conversions: [matched] covers. Read cost per conversion against [cap]. If it sits above the cap, cut. The tool does not invent incrementality or a 3x benchmark.

Hotel and retreat reuse

Swap ticket for ADR, food cost % for room cost %, and covers for room-nights. Upload booking-engine emails you own, not OTA guest lists. Direct-booking extras stay in direct bookings.

7-day actions

1. Day 1: Export last 4 weeks of average ticket, food cost %, and weekly covers. Fill [avg_ticket], [food_cost_pct], [contrib], [weekly_spend]. Set [floor] and [cap].
2. Day 2: List booking tools and whether they store email or phone. Create or confirm the Meta offline event set on the spending ad account.
3. Day 3: Fire pixel Purchase only on the reservation confirmation URL you own. Stop optimizing for landing-page views or profile visits.
4. Day 4: Upload last week's matched reservations and CRM POS checks. Keep order ids.
5. Day 5: Host-stand card: "Did you find us on Instagram or Google?" as a note, not as the ROI number.
6. Day 6: Open the social media ROI calculator with spend, attributed revenue, and matched covers.
7. Day 7: Run both gates. Scale only if matched covers beat the cap for two consecutive weeks. Kill ads that fail the optimistic floor. No-go if you are still optimizing for landing-page views or scaling because Ads Manager shows 3x. Partial go: optimistic floor passes and matching is thin, hold spend and fix the file for two weeks.

Ad copy still has to be claim-safe. Do not call the dining room eco-friendly in the ads you are measuring. Pair with LLM ad-copy guardrails.

Frequently asked questions

How should vegan restaurants measure Meta and Instagram ads ROI when most covers are walk-in or first-party reservations?

Set a contribution floor and a cost-per-cover cap from POS, upload offline conversions weekly from the booking tool or CRM, then open the social media ROI calculator with matched covers and POS-attributed revenue. Do not use Ads Manager ROAS or invent a 3x target.

Why does Meta ROAS not match the dining room?

Walk-ins and most first-party reservations never fire Purchase. Ads Manager can still credit diners who would have found you on Maps. Use POS contribution math for go/no-go and offline uploads to teach the algorithm.

What is a good ROAS for vegan restaurant ads?

There is no honest 3x rule for dining. Set cost-per-cover at or below contribution after food. If even spend divided by all covers exceeds contribution, the ads cannot be paying.

How do walk-ins get into the weekly upload?

Only if the guest is already in your CRM with email or phone. Do not invent emails. Unmatched walk-ins stay unattributed. Use them only in the optimistic gate (spend divided by all covers).

Can vegan hotels use the same worksheet?

Yes. Swap ticket for ADR, food cost for room cost, and covers for room-nights. Upload booking-engine emails you own, not OTA guest lists.