Case study · Restaurant performance data and analytics

Turning restaurant-market signals into relevant executive outreach

A restaurant brand can produce dozens of potentially interesting signals in a year. Detecting one is not the same as understanding why it matters. We built the interpretation layer between them.

Download detailed case study (PDF)PDF · 7 pages

Case study · Restaurant technology intelligence

Black Box Intelligence (BBI), restaurant performance benchmarks and market data.

6 campaign motions on one research system
10 decision steps before a message
The signals
New openings, expansion plans, LTOs, earnings calls, traffic and pricing pressures, guest-experience initiatives
The missing layer
Interpretation: what does the signal mean operationally, and can BBI help with that problem?
The constraint
Prompts grounded in BBI's actual data and capabilities, not a language model's general restaurant knowledge

The problem

A signal is not a reason

A company hired a new executive. A restaurant launched a limited-time offer. Traffic declined. An executive attended a webinar. All are signals; none automatically means the company is ready to buy.

Why generic personalization fell short

  • “I saw you're expanding into new locations…” plus “we help restaurant brands make better decisions with data” — each statement can be reasonable while the connection between them stays weak.
  • Early generated challenge-and-solution statements leaned on the model's general restaurant knowledge: plausible, without being grounded in BBI's actual data and capabilities.
  • The useful sequence ran through questions first: what does the signal mean operationally? Can BBI help with that problem? Which capability is relevant? Which executive function cares?

What we built

Interpretation before outreach

  1. Company and ICP qualificationU.S. restaurant fit first — a global hospitality company's international activity was not automatically relevant. HubSpot lookups added internal CRM context.Define
  2. Restaurant initiative researchNew-unit expansion, multi-unit growth, venue formats, guest-experience initiatives — initiatives that could create a legitimate performance question.Qualify
  3. Match the initiative to a problem BBI can help solveThe event is not the pitch: market-evaluation questions, benchmarks, guest sentiment and peer comparisons turned an expansion into a business problem.Qualify
  4. Capability-grounded promptsResearch logic grounded in what BBI actually measures: performance benchmarks, sales and traffic analysis, guest sentiment, workforce data, peer comparison.Contact
  5. Role-aware interpretationThe same earnings evidence reads differently to finance, operations, marketing and guest experience — the angle changed by recipient function.Launch

Six campaign motions reused the research: strategic initiatives, state-specific geographic intelligence (chosen by a fixed hierarchy — most brand locations, then HQ, then contact's state), LTOs, earnings calls, webinar follow-up and evergreen outreach. Deterministic routing applied AI only when a prior step was unavailable.

Production realities

Judgment stayed in the loop

  • A real initiative could still be the wrong initiative: opening a headquarters doesn't make restaurant-performance data relevant. The system chose the initiative BBI could help with, not the newest one found.
  • One signal could have several interpretations — HR, operations, marketing or finance — and recipient role shaped the reasoning.
  • Earnings calls contained many metrics; only strategically relevant numbers survived into outreach. Geography could not be guessed: fallback logic made the state decision deterministic.

Human review and iterative prompt design added value here. By the time a message was written, the account had been qualified, its signals tested for relevance, and the evidence interpreted for a specific executive.

The signal was only the beginning. Research created the reason.
Next caseGoGather: turning past event targets into a prospecting system →

Related reading

A practical guide to the research and system-design decisions behind work like this.

B2B buying signals vs. buying intent

Let's discuss your GTM challenge.

A short, no-obligation introductory conversation.

Schedule a conversation