For a small restaurant that misses calls in service, often. Yelp Host starts at $249 a month by Yelp's own list price ($199 if you use Yelp Guest Manager), so at a 30% margin it pays for itself if it wins about six extra table bookings or 25 extra takeout orders a month. It needs an accurate Yelp listing, and for orders, a connected POS.
It is worth less than it looks if your calls are mostly staff-answered, your Yelp page is thin, or your phone orders are few. The honest test is not features but a break-even sum on your own missed calls, run before the free trial, then checked against what the trial actually catches. Most of the cost of getting it wrong sits in the listing, not the software.
What Yelp Host costs, and why old prices are wrong
Older third-party write-ups quote lower launch prices, and they are out of date. Yelp's 28 July 2026 news post says Yelp Host starts at $249 a month, with $199 for Guest Manager customers, and the restaurant product page shows a $249 to $399 range in its comparison table. The page does not explain what moves a restaurant between those figures, and it does not state a call limit. So when you start, ask your rep in writing which tier you are on, what is included in call volume, and whether the price is fixed for the term.
The same page says you can get Yelp Host free for up to 30 days, subject to extra fees and terms. A free trial is the right way to learn what you need to know, but only if you set it up well. Run it across a full week including your busiest service. A trial on a Tuesday in February tells you very little about a Friday in December.
Annualised, $249 a month is $2,988 a year, and $399 is $4,788. Those are the numbers to put against your own missed-call losses. Yelp's own comparison table puts Slang at $399 to $599 a month and Revmo at $0.59 a conversation, which gives you two reference points from the vendor's side; check each rival's own page before you trust a competitor's summary of it.
The break-even sum, with your own numbers
Take an illustrative 40-seat restaurant that is open six evenings and one lunch a week. Staff are slammed from 6.30pm to 9pm, and the phone rings through. Suppose they log missed calls for one week and find 30. They call back what they can, but a fair share of callers have booked elsewhere by then.
- Say 30 missed calls a week, of which a third (10) are booking requests, a third (10) are takeout orders and a third (10) are questions such as opening hours or parking.
- If Yelp Host can convert half of the booking requests into held tables and half of the orders into taken orders, that is 5 bookings and 5 orders a week.
- Suppose the average booking is a party of three spending $45 a head, so $135, and the average takeout order is $32. That is 5 times $135 plus 5 times $32, or $835 a week.
- Not every booking turns up, and not all of it is extra business because some callers would have rung back. Apply a haircut of 50%: about $415 a week, or around $1,800 a month in revenue.
- At a 30% food and drink margin, the contribution is roughly $540 a month. Against $249, the phone service looks worthwhile. Against $399, still positive, though thinner.
Run it again with half the missed calls and the same assumptions: contribution falls to about $270, barely above the lowest price. That is the threshold. Below about 15 missed calls a week that could realistically convert, Yelp Host is a close call. Above 30, it is an easy yes. The figures are illustrative, so replace each with your own.
How to count missed calls properly: for one week, have whoever closes note each unanswered number from the call log, ring a sample back, and write down what they wanted. Your phone provider's call history is enough, and a spreadsheet with three columns (time, what they wanted, outcome) does the rest.
What Yelp Host needs from your Yelp listing
Yelp's restaurant page says Host pulls hours, location and business information from your Yelp page. It is described as pre-trained on Yelp business information and works from the start. That makes your listing the first thing it learns from, and a stale listing becomes a confidently wrong phone answer.
Before you start a trial, go through this list:
- Hours, including holidays and kitchen close. If the kitchen stops at 9pm but the bar stays open till 11, say so. Special hours for bank holidays are the most common error.
- Address and parking. Callers ask where to park more than anything else. Put the answer on the listing and in Host's custom answers.
- Menu and dietary notes. Yelp says you can customise menu and dietary details, amenities and special offers. Write plainly what contains nuts, what can be made vegan and what cannot.
- Policies. Dogs, high chairs, large-group deposits, corkage, cancellation. These are the questions where a wrong answer causes a row at the door.
- The designated transfer number. Yelp's page says transfers need a designated number set in your Yelp for Business account. Set it to a phone that someone actually carries in service, not the office landline.
A before and after of one answer
Here is an illustrative case. The listing says "Open daily 12 to 10". The kitchen in fact closes at 9, and on Sundays you close at 8. A caller at 9.20pm asks, "Can I still get food?"
Before (listing as it stands):
"Yes, we are open until 10pm tonight. Would you like to book a table?"
After (hours and kitchen close fixed, Sunday exception added):
"The kitchen closed at 9pm tonight, but the bar is open until 11. I can put you through to a team member if you would like to ask about a late snack."The software did not get smarter between the two. The listing got more accurate. Most disappointments with phone AI in restaurants trace back to source information, and fixing it costs an hour.
Reservations, waitlist and takeout: what each needs
Reservations
Yelp says Host can use both your online and in-house availability through Yelp Reservations, and that it integrates with OpenTable, where callers can book, change or cancel by phone, with bookings syncing in real time. The OpenTable link covers OpenTable's network in a limited set of markets, so check yours. Yelp's restaurant page also says Host books reservations and waitlist entries with your booking platform. If your tables live in a different system, ask which are supported before you rely on this.
Waitlist
The product page says Host handles waitlist entries in real time once your booking platform is connected. Beyond that I found no details on how it quotes waiting times, so test it: ring at a busy moment and see whether the quoted wait matches your host stand.
Takeout payment links
For orders, Yelp's announcement says Host takes pickup orders by phone, pulling your menu through Stream (Yelp's connector) from your POS, with Toast, Square and Clover named, "and others". The customer can pay at pickup or receive a text with a link to prepay. Confirmed orders flow into the POS and kitchen display with no re-keying. You can set takeout hours, pause ordering, mark items sold out and set order size limits. Yelp states there are no commission fees, unlike delivery apps, though your card processor still takes its usual cut. The announcement covers pickup orders only and does not describe delivery.
The practical consequence is that your POS menu must be clean. Duplicate items, "Special 2" entries and modifiers that make no sense to a caller become spoken confusion. Do the cleanup in the POS first. The delivery app commission tutorial shows what the commission-free route is worth against marketplace orders.
Transfers to staff: set the rules deliberately
Yelp's post says you can set multiple handoff numbers, so vendor calls go to management and event requests go to an events coordinator. A single transfer number is the default, and the sensible rules are short:
- Always transfer: allergy questions beyond what you have written down, complaints, anything about a previous visit, press or suppliers.
- Never transfer when no one can answer: if nobody carries the phone during a rush, transfers just recreate the missed call. Better to let Host take a message or send a text link.
- Time-limited transfers: if the page lets you choose when Host answers (it can cover 24/7 or only when you need extra coverage), choose only the busy windows first.
Take an illustrative pizzeria where the owner finds that 80% of Friday-night calls are orders and 20% are complaints and supplier chasers. The owner leaves order calls to Host and sets supplier and complaint calls to go straight to a mobile. Friday missed calls drop because the only calls reaching a human are those that need one.
How it compares with doing nothing or with a cheaper fix
Before paying $249, consider the free and cheap alternatives, because Yelp Host is not the only way to stop missing calls.
| Option | What it fixes | What it leaves |
|---|---|---|
| Keep the listing right, add an online booking link | Callers who would book online anyway | Callers who insist on talking |
| Voicemail with a text-back link | Some lost callers, for little or no cost | Impatient callers, orders in a hurry |
| Yelp Host from $249 a month | Answering, reservations, takeout orders, language detection | Complaints, odd requests, unreliable POS menus |
| Another restaurant phone AI (Slang, Maple, Popmenu and others; see the restaurant receptionist comparison) | Similar, with different pricing and integrations | Depends on the vendor |
If a text-back link and a better listing recover most of what you lose, you do not need a $249 service. If the lost calls are takeout orders while you are chopping, a link will not help and Host looks better. The phone answering cost tutorial sets out the wider price picture across vendors.
Five ways it can go wrong, and how you would spot them
- A wrong answer on policy. Spot it by reading call transcripts and summaries daily during the trial. Fix the source (listing or custom answer), not the symptom.
- An order lands in the POS with a wrong item. Spot it by checking the first twenty orders against the caller's recording. Look for substitutions and modifier mix-ups.
- Double-booked tables. Spot it by checking where reservations land. If the same table can be booked online, by phone and in person from three systems, you will collide.
- Regulars hanging up. Spot it by asking regulars what they thought. If locals dislike the voice, route known numbers straight to staff or adjust the greeting, which Yelp says you can customise.
- Calls that should be a person. Spot it by listing the topics where Host transferred and the ones where it should have. Tighten rules weekly.
A two-week trial that gives a clear answer
Record the baseline first. For one week before starting, log missed calls, order calls and booking calls, as in the sum above. Then:
- Days 1 to 3: fix the listing, clean the POS menu, set transfer rules, write custom answers for parking, dogs, allergens and large groups.
- Days 4 to 7: run it only in your busy window. Read every call summary.
- Days 8 to 14: widen to all hours if the first window was clean. Count orders and bookings it handled and compare with the baseline.
- Decision: take the monthly price, divide by the average contribution of a booking or order, and see how many extra it must win. If the trial produced twice that many, buy it. If it produced about the same, treat it as a coin toss and look at cheaper options.
Keep a note of what happens when you cancel too. Ask whether your dedicated number is yours, whether you can forward your existing line instead, and what happens to call data. The exit terms tutorial lists the questions.
Reading a call summary: a sample and what to correct
During the trial, the most valuable ten minutes of your day is reading what Host did. Here is an illustrative summary of a takeout call, with the points an owner would check:
Caller asked for two margherita pizzas, one with no basil, and a garlic bread.
Pickup at 7.15pm. Name: [first name]. Payment: text link sent.
Outcome: order sent to POS. Transfer: none.Check three things. First, did the "no basil" note reach the kitchen ticket, or only the summary? Second, was 7.15pm realistic given the tickets already on the rail? Third, did the payment link get paid before the pizza was made? If the answers are yes, yes and no, you have found a rule to set: only fire prepaid orders, or take pay-at-pickup orders only before 6pm. Repeat that review for ten calls and you will know the failure patterns of your own menu better than any vendor demo could tell you.
Another example is a booking call. A caller asks for a table for eight on Saturday. Your policy says parties over six need a deposit. If Host books it without mentioning the deposit, add the rule to its custom answers, and consider routing parties over six to a person until you trust it.
Briefing the team so it does not become a mystery box
Staff who do not know what Host does will either distrust it or blame it. A short briefing prevents both. Tell the team that calls to the published number may now be answered by an assistant, that orders arrive in the POS marked as phone orders, and that transfers will ring a named phone. Put the rule for handling a transferred call on a card by the till: answer in your own name, say you have been passed a call, and never ask the caller to repeat everything, because Host has already collected the basics.
One more habit: a weekly ten-minute review of five transcripts, where each person brings one call that went well and one that went badly. The notes feed straight back into the listing and the custom answers, which is where most improvements live.
Prepaid orders and the quiet-hour effect
Prepaying by text link changes two numbers that owners rarely measure. No-shows on takeout fall, because the customer has paid. And order timing becomes easier to manage, because you can see paid orders only. As a quick sum, a restaurant taking 15 phone orders a week at $32 with a 6% no-show rate loses about $29 a week to orders nobody collects, roughly $125 a month. If prepaid links remove most of that, the saving covers about half of the lowest monthly price on its own. The figures are illustrative, but the logic holds: count your own uncollected orders before you decide the service has no pay-back.
The verdict by restaurant type
A busy table-service restaurant with a Yelp listing it already maintains: likely worth a trial. Phone peaks coincide with service, so the cost of a missed call is highest exactly when staff cannot answer.
A takeaway or counter-service place with a clean POS: worth it if phone orders are meaningful. Order-taking with no commission is the strongest argument, and the break-even is easy to count in orders.
A quiet neighbourhood restaurant where the owner answers the phone: probably not yet. The numbers will not clear $249 a month unless calls are being missed.
A restaurant with no active Yelp presence: fix that first, because Host learns from the listing. Without it, you are paying a service whose source material is empty.
If you would rather size the problem with someone before spending, that is what my AI implementation consultation covers. Otherwise, run the log, do the sum and use the trial.
Yelp Host questions owners ask
How much does Yelp Host cost?
Yelp's own pages show Yelp Host starting at $249 a month, or $199 a month for Yelp Guest Manager customers, with a comparison table running to $399. Older third-party write-ups quote lower launch figures, so treat them as out of date and confirm your tier with Yelp in writing.
Is there a free trial?
Yelp's restaurant page says you can get Yelp Host free for up to 30 days, with additional fees and terms applying. Use the trial on your busiest week, not a quiet one, so the call volume you test reflects real life.
Can it take takeout orders?
Yes, according to Yelp's announcement. It takes pickup orders by phone, pulls your menu from the POS, and the customer pays at pickup or gets a text link to prepay. Yelp says there is no commission, though your card processor still charges its own fees.
Does it need a Yelp listing?
Yelp's page says Host pulls hours, location and business information from your Yelp page. A thin or out-of-date listing means thin or out-of-date answers, so tidy it before the trial starts.
Further reads
- Slang.ai for Restaurants: Is It Worth It for an Independent? — The pricier rival Yelp's own table compares Host with.
- Maple AI for Restaurants: Voice or Pro Plan, and Who Needs Which — A flat-priced phone AI to test Host's price against.
- Can AI Answer the Phone and Take Restaurant Reservations? — How AI handles reservations over the phone.
- AI Restaurant Order Taking: Who Takes the Order, Who Texts a Link — Who takes the order and who only texts a link.
- Popmenu AI Answering: Does It Take Orders or Only Send Links? — A competitor that sends links rather than taking orders.
- Should an AI Chatbot Answer Allergen Questions for Your Restaurant? — Which allergen questions an AI should never answer alone.
- Pizza Phone Ordering System: What AI Costs per Month — Per-minute, flat-fee and per-customer AI phone pricing for a pizza shop, with a missed-call payback sum and the extras that never show on the price page.
- SevenRooms Voice AI: What It Does and Who It Suits — What SevenRooms Voice AI does, why you must ask for a quote, and how it compares with Yelp Host, Slang and Popmenu.
- Restaurant AI Voice Agent: 12 Test Calls to Run Before Go-Live — Twelve scripted calls with pass marks, blocker or fix labels and a go or no-go rule, so you find the failures before a guest does.
- Restaurant Waitlist Software: Where AI Helps at the Door — AI can quote waits by party size, text when a table is ready and answer waitlist calls. Here is what it still needs a host at the door to do.
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: Yelp's Yelp Host restaurant product page and Yelp's 28 July 2026 news post on OpenTable reservations and takeout ordering, read in October 2026.