DEV Community

naoanao
naoanao

Posted on

3 hrs/week on restaurant marketing 30 min. Here's the exact system I built.

Every restaurant owner I know has the same weekly problem.

You know you should post on Instagram. Reply to Google reviews. Run a promotion. But every week you stare at a blank screen: "What should I actually do?"

I ran my burger shop in Japan with this exact problem for years. 3+ hours every Sunday just figuring out what to do.

The fix: local competitor intel + AI = 3 specific actions per week

After 6 months of building and testing, here's what actually works:

Step 1: Scan what competitors in your area did THIS week (not generic industry trends)

Step 2: Find the gap — what are they doing that you're not?

Step 3: Generate 3 specific, copy-ready actions from that intel

That's it. Three actions. You can actually do three things.

My current weekly routine

  • Mon 8am: Check 3 auto-generated actions (2 min)
  • Mon 8:30am: Schedule the Instagram post (10 min)
  • Tue: Reply to Google reviews using the template (5 min)
  • Thu: Run the promotion idea if it fits (15 min)

Total: ~30 minutes. Before: 3+ hours.

I built Growl to automate this

Growl does this automatically for any restaurant owner.

Answer 5 questions about your restaurant once. Every week it:

  • Runs live competitor research in your area (Tavily real-time search)
  • Generates 3 specific copy-ready marketing actions
  • Delivers: Instagram post copy, Google review reply template, local promo idea with timing

Free to try. No signup required.

The technical bit (for builders)

Stack: Next.js + Groq LLaMA 3.3 70B (~500 tok/s) + Tavily live search + Cloudflare D1

The key insight: LLMs are useless for local marketing without current, local data. Generic prompts give generic advice. Grounding with Tavily real-time search is what makes this work.


Try it free → growl-app.vercel.app

Day 363 of building solo. $0 revenue — keeping it honest.

restaurant #marketing #ai #buildinpublic

Top comments (0)