Earlier this year, we found something incredible at Salad House (and ive tested it with other brands I'm working with too.)
Guests who tried our Cauliflower Bites were far more likely to become high-frequency guests. So we tested it. We found guests who were overdue for a visit and sent them an offer for free Cauliflower Bites. From email alone, that campaign drove nearly $300,000 in sales, and 23% of those guests made another purchase within a couple of weeks. It worked so well that it's now an always-on automation.
Then we did what you should always do after a win. We asked why.
THE NEXT LAYER
This fall we rebuilt our guest data and stacked more analysis on top of the sales: how guests order, whether they're in Rewards, how fast they come back, and what they say about us. The story got richer. Cauliflower Bites guests lean toward ordering on the app, and app guests are some of our most loyal guests.
So the Bites weren't just a great item. They were pointing us to a bigger habit worth building.
Now we have two levers instead of one: a craveable item that brings people back, and an ordering habit that keeps them coming.
One data source will give you a confident answer. Multiple data sources will give you the whole picture.
WHY ONE LENS ISN'T ENOUGH
Understanding your best guests is how you reverse engineer your entire funnel. Who they are, what they ordered first, how they ordered it, how fast they came back, and what they love about you. Know that, and you know exactly what to build.
But every system you own only sees part of that guest.
Your POS knows what they bought. Your loyalty platform knows who they are and how often they come. Your review and sentiment platform knows what they love and what drives them nuts. Social shows you what gets a reaction, which isn't always what gets a return visit.
Look at any one of those on its own and you get a story. Stack them and you get the novel.
Here's the headline from ours. A small slice of our guests drives a huge share of our revenue, our future regulars build the habit in their first month, and our biggest opportunity wasn't a menu item at all. It was how and where we invite guests into the relationship.
HOW TO DO IT
1. Pull from every source you have.
POS for what they buy.
CDP or CRM to centralize guest data and power your communications.
Loyalty for who they are and how often they come.
Reviews and surveys for why they love you.
Social media for what's landing for the brand and what's not.
Your team for the context only you know, like openings, recalls and operational changes. Data without context will mislead you.
You don't need a data warehouse. You need the exports and a good set of questions.
2. Tie it to a real guest. Every order needs a guest ID attached, or you're analyzing a crowd instead of people. Find out what share of your orders you can actually identify. If most are anonymous, fixing that comes first.
3. Build your ladder. Group guests by visits: 10 or more, 5 to 9, 3 to 4, 2, and 1. For each rung, look at what share of guests they are and what share of revenue they drive. Prepare to be surprised.
4. Compare each rung to the one above it. What do your 5-to-9 guests do differently from your 10+ guests? How fast did each group make their first few visits? How did they order? What did they say about you? The answers are usually simpler than you'd expect.
5. Check yourself before you bet on it. Split every pattern by channel, because an app pattern can look like a menu pattern. Look at guests' first 90 days, not their lifetime, so you find the behaviors that created loyalty and not the ones that came after. And make sure the why (sentiment) matches the what (purchases).
6. Test one thing and scale what works. Pick one pattern and build an offer or a journey around it. Hold back a small group that doesn't get it, so you know the result is real. Then turn it into an always-on automation.
THE TECH
People always ask what tech we use. For Salad House:
We benchmark against industry research, like Bloom Intelligence. Our team adds the context only we know. Claude does the heavy lifting on the analysis. A year ago this would have taken a data team. Now it takes a good set of questions.
Get your good guests to be more like your best guests. Just make sure you know what actually makes them your best.
Over the next few weeks, I'm going deep on each of these steps. Send me any questions you have. But in the meantime, here's my question for you: what's one thing you believe about your best guests that you've never checked against a second source?
Do you need help figuring out who your best guests really are, or using AI to get there faster? Send me an email at [email protected].
- Rev Ciancio
WHAT DOES REV DO?
I help restaurants build guest marketing programs.
I help hospitality tech companies with lead generation and content marketing.
