I run a 22-seat Italian restaurant in a midwestern suburb. We have been open for eight years. Yelp is the platform our customers actually use, more than Google in our area. We had 416 reviews and a 3.9 average when I started this experiment. The 3.9 was killing us. People stop reading at four stars.
I had a long-standing rule against replying to reviews. I had watched too many restaurant owners get into screaming matches with one-star reviewers on Yelp and look like maniacs. My rule was: do not respond, do not engage, do not make it worse. I decided to test the opposite hypothesis for six months.
The rules I set for myself
Three rules, written on a sticky note next to my laptop. One: I respond to every review, positive or negative, within 48 hours. Two: I respond from a place of curiosity, not defense, even when the review is unfair. Three: I never write the reply when I am angry. I write a draft, walk away for an hour, and come back.
Month one: the one-star reviews almost broke me
The first one-star review came in on day three. The customer claimed our pasta was undercooked and the waiter was rude. I knew the waiter. I knew the pasta. I drafted a defensive reply explaining the cooking process. I walked away. I came back. I deleted the draft. I wrote a different one: 'I am sorry your meal did not meet your expectations. I would like to understand what happened. If you are willing, please email me directly at the address below.'
She emailed me. We exchanged six messages over a week. It turned out she had been served the wrong dish, which was not undercooked but was someone else's order, and she had thought it was the dish she ordered. The waiter had been short with her because his shift had run six hours past when his kid's school called about a fever. She did not change her review at the time but she said she would come back.
She came back two months later, ate her actual order, and updated her review to four stars with a note explaining what had happened. That was the moment I knew the experiment was working.
The five-star replies that surprised me most
I had assumed responding to five-star reviews was a nice-to-have. I was wrong. Five-star reviewers turn into the most loyal customers when you reply. I started writing personal, specific responses. I would mention what they had ordered, ask about their friend they mentioned in the review, and sometimes tell them a tiny piece of behind-the-scenes context they had not known.
Within two months I could trace at least 30 visits a month back to customers who had left five-star reviews, gotten a personal reply, and come back specifically because they felt seen. Three of them brought new groups of four to six people for birthdays.
The lawsuit threats
I had two reviewers threaten to sue me for libel after I replied to their reviews. Both were one-star reviews that contained factual claims I knew to be false. In one case the reviewer claimed we had served their child a dish containing peanuts despite a warning. We had never served them a dish with peanuts. I had the order receipt and the kitchen ticket.
My reply did not call them a liar. It said: 'I take peanut allergies very seriously and our kitchen process flags any peanut-containing dish. I have pulled the receipt from your visit on [date] and the dishes ordered did not contain peanuts. If you believe there was cross-contamination, I would like to talk through the kitchen process with you.' The reviewer accused me of harassment. I called a lawyer. The lawyer said my reply was professional and factual and there was no case. The reviewer never responded again. The review stayed up. Several months later a different customer left a five-star review that specifically said they had felt safe with their nut allergy at our restaurant.
Month three: the people who never review you
About 10 weeks in I noticed something I did not expect. New customers were mentioning my Yelp replies. Not the reviews themselves, but my responses. They had read through 20 or 30 of my replies before coming in and decided they wanted to support the place. One couple told me they had read every single response over the previous year before booking their anniversary dinner.
This was the part of the experiment no one had told me about. The replies were not really for the reviewers. They were for the next 5,000 people who would read those reviews and replies before deciding whether to come in. Every reply was a billboard.
Month four and five: the slow drift upward
Between month one and month five my Yelp rating moved from 3.9 to 4.3. The number of new five-star reviews per month doubled. The negative reviews did not stop, but the percentage of negative reviews that updated after my reply was 22 percent. About one in five.
Month six: the final accounting
- Reviews replied to: 182.
- Average response time: 14 hours.
- Negative reviews that updated after my reply: 11 of 49 (22 percent).
- Five-star reviewers who came back within 60 days of my reply: 38 of 113 (33 percent).
- Average rating: 3.9 → 4.3.
- Lawsuit threats: 2. Lawsuits actually filed: 0.
- Hours per week spent: roughly 4.
What I would tell another restaurant owner
If you do nothing else in your business this quarter, set aside two hours per week to reply to reviews. Reply to the good ones with specifics. Reply to the bad ones with curiosity, never defense. Walk away before you hit send if you feel any heat in your chest. And remember the replies are not for the reviewer. They are for the next thousand people who will read the page.
Soft note: I now use Social Perks to run this. The 14-day free trial was enough time to know it was the system I had been duct-taping together for years.