Restaurant CRM is really guest management, and the market splits at a hard line. Reservation-led platforms publish prices from about $149 a month. Full guest data platforms start nearer $795 a month per location, with implementation on top.
Between those two sits the question that decides your spend: do you need to know who the guest is, or only that a table is booked?
The term itself causes confusion. Customer relationship management in a restaurant is not the sales pipeline software the phrase means everywhere else. There is no deal to track and no salesperson to prompt. What you are buying is a unified guest profile: one record per person, holding what they eat, where they sit, when they last came and what they spent, joined across every visit and, if you have several sites, across all of them.
Most restaurants overbuy here, because a demo of a full guest profile with allergy notes, spend history and visit frequency is genuinely impressive. Whether you can act on it depends on things that have nothing to do with software. This page covers the real prices, the four capabilities you are choosing between, and how to tell which side of the line you are on.
What the platforms cost
| Platform | Type | Published price | Note |
|---|---|---|---|
| OpenTable | Reservations, with guest data | $149 / $299 / $499 per month | Plus roughly $1 to $1.50 per cover on network reservations, which is the line that grows. |
| Toast | Point of sale, CRM bundled higher up | From about $69 per month | That entry price is the till, not the guest platform. CRM sits in higher tiers. |
| SevenRooms | Full guest data platform | About $795 per month, single site | Implementation commonly $1,500 to $3,000. Multi-site moves toward custom pricing. |
| Punchh | Loyalty and marketing | About $400 to $2,000+ per month | Per brand, often structured per location or per active loyalty member. |
| Asyntai | Guest answering | $0 / $39 / $139 / $449 | 100, 2,500, 15,000 and 50,000 messages a month. Answers guest questions on your website. |
We make Asyntai, so read that row as a disclosure. It is the cheapest line on the table and it covers a job the others leave to your staff: answering the phone and website questions that arrive before anyone becomes a guest, at any hour and in any language. Prices for the other four come from published 2026 comparisons rather than first-party rate cards in every case; confirm each with the vendor.
The per-cover charge is the one to model
Reservation platforms commonly charge a monthly fee plus a fee per diner seated through their network. The monthly fee is the number in the marketing. The per-cover fee is the number on your bill.
Two consequences worth acting on. First, ask for the split between network covers and your own direct bookings, because only one of them carries the fee. Second, every booking you move to your own website is a fee you do not pay, which makes your direct booking channel worth real money rather than being a nice-to-have.
Work out what a network cover costs you, then work out what a direct booking is worth. That number funds everything else.
The four capabilities you are choosing between
Vendors bundle these differently, and the labels are unhelpfully similar. Sort them and the shortlist gets short.
1. Reservations and table management
Taking the booking, seating the room, managing waitlists and covers. Every platform here does this competently. It is table stakes and it should not decide your purchase.
2. The guest profile
Knowing that this is the couple who come every third Friday, sit at table 12, drink Barolo, and one of them is allergic to shellfish. This is what people mean by restaurant CRM, and it is the capability that costs money.
The honest question: can your team act on it? A profile note only matters if the host reads it before seating and the server sees it before taking the order. If your service does not have that step, you are buying a database nobody opens.
3. Marketing and loyalty
Campaigns, offers, points, win-back messages to guests who have not visited in ninety days. Distinct from the profile, frequently sold separately, and the place where per-member pricing appears.
4. Spend history joined to the guest
This needs your point of sale and your booking system to be talking to each other. It is the hardest one to get working and the most valuable when it does, because it turns a visit count into a revenue figure per guest.
Ask specifically how the join happens. If it depends on staff manually matching a check to a booking, it will be half accurate within a month.
Single site or multi-location changes the answer
One restaurant
- Your team already recognises regulars. The software adds less than the demo suggests.
- Reservation-led at $149 to $299 is usually enough.
- Spend the difference on the direct booking channel instead.
- Watch the per-cover fee; it will be your largest line.
Several sites, one brand
- A guest who visits three of your sites should be one record, not three.
- This is the genuine case for a central guest database.
- Ask how records merge across sites, and what happens to duplicates.
- Price implementation separately; it is not small.
The multi-location case is where the expensive platforms earn their price, and it is exactly the case where the cheap ones fall down. If you have one site and a good host, be honest about which column you are in.
The queue nobody prices
Before any of this software touches a guest, somebody has already answered a question. Restaurants take a remarkable volume of the same enquiries, and almost none of it is in a CRM.
- Are you open on Monday, and what time do you stop serving?
- Do you have anything gluten free, or a vegan option?
- Can you seat eight people on Saturday?
- Is there parking, and are you accessible?
- Do you take dogs?
- Do you do set menus for groups, and what do they cost?
These arrive by phone during service, by email overnight, and through social messages at midnight. Every one of them has a published answer on your own website. Handling them costs you a member of staff mid-shift, and missing them costs you a booking.
Answer the 11pm question about Saturday
Asyntai reads your existing website, menus, opening hours, allergen information, group bookings and access, and answers guests from it at any hour, in their own language. The booking still happens in your reservation system.
Try it on your restaurant siteGuest preferences: what is worth tracking
Every guest management platform will let you track dozens of fields. Almost none of it changes an evening. Six things do, and a restaurant that records only these six will outperform one that records forty and reads none.
| What to track | Captured when | Read when | Why it earns its place |
|---|---|---|---|
| Allergies and dietary needs | At booking, confirmed at the door | Pre-shift, and at the pass | Safety first, and a guest who is not asked twice feels remembered. |
| Table preference | Observed, not asked | When seating | The cheapest way to make a regular feel like one. |
| Occasion | At booking | Pre-shift | Anniversaries and birthdays are the highest-return note in the system. |
| Wine or drink preference | After the order | When taking drinks | Raises spend without anyone selling. |
| Last visit and frequency | Automatically | Monthly review | Tells you which regulars have quietly stopped coming. |
| Spend per visit | From the point of sale | Monthly review | Turns a visit count into a number you can act on. |
Notice the middle column. A preference that nobody reads at the right moment is not tracked, it is stored. The pre-shift briefing and the moment of seating are the only two points where a note changes anything, so design your fields around those two moments and delete the rest.
Guest retention strategies that use the data
Buying the platform is the easy half. Four strategies actually use what it collects, and they are worth more than any campaign feature.
The ninety-day lapse list
Once a month, pull the guests who used to come every few weeks and have not been in for ninety days. That list is short, specific and actionable. A short personal message from the manager, not a marketing blast, brings a meaningful share of them back. This one report justifies most of what a guest platform costs.
Second-visit conversion
The gap between a first visit and a second is where most guests are lost, and it is measurable. Track what share of first-time guests return within ninety days. If you improve one number this year, make it this one, because a second visit predicts a tenth far better than a first one does.
The occasion calendar
Anniversaries and birthdays recorded this year are bookings next year, and almost nobody works this list. A message a fortnight before the date, offering the same table, converts unusually well because the guest was probably going to book something anyway.
Improving the guest experience with what you already know
The most valuable use of a guest database is not marketing at all. It is the host greeting somebody by name, the server knowing not to recommend the shellfish, and the table by the window being free because somebody noted that they like it. None of that is a campaign, and all of it comes from the same six fields above.
The guest platform pays for itself at the door, not in the mailout.
Five questions before you sign
- What is the per-cover fee, and on which bookings does it apply? Get the split between network and direct in writing, then model it at your real cover count.
- What does implementation cost, and how long does it take? On the guest platforms this is frequently thousands and several weeks. It belongs in the year one number.
- How does the guest record join to spend? Ask for a live demonstration of the point of sale connection, not a logo on an integrations page.
- Who owns the guest data, and can we export it? This is your list. Get the export format and the notice period in writing before you put ten thousand guests into it.
- What happens to duplicate guests across sites? If you have more than one location, this single answer decides whether the central database is real or theoretical.
Making it actually work in service
Guest platforms fail in restaurants for the same reason CRMs fail everywhere: the person entering the data is not the person who benefits from it, and service is busy.
Four habits separate the restaurants where this works from the ones where the database rots:
- One note field, used consistently. Not six custom fields nobody fills in. Allergies, preference, occasion. That is enough to change a guest's evening.
- The host reads it at the door. If the pre-shift routine does not include looking at tonight's book, no software will help.
- Capture at the point of pleasure. The moment to note that someone loved a dish is while they are telling you, not at the end of the night from memory.
- Review it monthly, not never. Ten minutes on which regulars have stopped coming is worth more than any automated campaign.
None of that is a feature you can buy, which is why two restaurants on the same platform get completely different results from it.
What to do next
- Count your covers and model the per-cover fee. On a reservation platform this is your real bill, and it is usually bigger than the plan.
- Be honest about whether you can act on a guest profile. If your service has no step where somebody reads it, buy reservations and stop there.
- Push bookings direct. Every direct booking avoids a network cover fee. That saving funds better software than the software you are avoiding it with.
- Handle the enquiry queue separately. It is high volume, out of hours, and entirely answerable from your own website. It is also the cheapest problem on this page to fix.
Sources
Platform prices for OpenTable, Toast, SevenRooms and Punchh, including the per-cover fee range and SevenRooms implementation costs, come from published 2026 restaurant software comparisons rather than first-party rate cards in every case. Confirm each figure with the vendor before you budget. Cover counts, enquiry volumes and staff time in the examples are illustrative, not measurements from Asyntai customers.