Discussion
  • Local SEO

Independent restaurants: what has moved bookings from search most: GBP menus and photos, reservation platform listings, or the website?

Published 1 reply

Discussion starter: a local SEO question, posted with an answer from Niraj Raut to open the thread. If you have dealt with this on a site, reply with what you saw, especially where it differs.

Disclosure: Niraj Raut, who posted this answer, runs the SEO consultancy linked at the end of it.

1 reply

  • admin
    Admin
    Original poster

    The lever that tends to move search bookings is not one of the three on its own. It is whether someone who finds you on Google can book in a tap or two. Diners who don’t know the restaurant yet tend to choose on the Business Profile in Maps and the local pack, so a reservation platform matters for search mainly when it puts a working booking button on that profile. Menus and photos work on choice: they decide whether people who see the profile pick you over the place down the road, and menu text can help you match dish and dietary searches. The website does its best work for people who already know your name, and for bookings that don’t fit a table for two: groups, functions, set menus and events.

    The qualification is that the answer for your restaurant is measurable, and cheaply. Bookings are countable, and a reservation system typically records a source for each one. Public data that isolates these three levers for independent restaurants is thin, and vendor-published figures deserve caution, so your own booking data is better evidence than anyone’s benchmark.

    How a search turns into a booking

    Search: “thai near me”, “[restaurant name] menu”, or “private dining for 20 in [suburb]”
    ↓
    Where the choice happens: the Business Profile in Maps or the local pack, your website, or AI Mode
    ↓
    What decides it: photos, menu, reviews, price level, opening hours, availability
    ↓
    Booking action: the booking button on the profile (through a reservation partner), the widget on your site, or a phone call
    ↓
    The booking lands in the reservation system with a source

    Three kinds of search behave differently. Discovery searches (“italian near me”, “brunch [suburb]”, “open now”) are decided in the map results, where proximity, relevance and prominence decide who appears, and the organic results below tend to be guides and aggregator lists rather than restaurant websites. Brand searches (“[name]”, “[name] menu”, “[name] booking”) show your profile and your website together, and the booking happens on whichever path is shorter. Occasion searches (“degustation”, “private dining room”, “gluten-free set menu”, “Christmas party venue”) are where a restaurant’s own pages can rank, and those bookings are often larger.

    Conversational booking adds a fourth path. Google’s AI Mode can take a request with time, party size, cuisine and mood, then look through multiple reservation platforms (with local partnerships) and websites to find bookable options. It started in the US, and in April 2026 Google began rolling it out to Australia, Canada, Hong Kong, India, New Zealand, Singapore, South Africa and the UK, with no Labs opt-in required. That report listed eight markets and no partner platforms, so check current availability in your market and ask your provider whether it is a partner. The practical point: a restaurant that can’t be booked online, through a platform or its own site, has nothing for that feature to book.

    What each lever actually changes

    Lever What it changes Where you see it When it is the priority The trap
    Business Profile menu (menu editor or menu link) Choice, and matching for dish, cuisine and dietary searches Profile interactions; the menu section of the profile Your menu is a reason to choose you: a specialist cuisine, dietary options, a price point Old prices, or a menu that exists only as a photo or PDF
    Photos Choice: what the room, the dishes and the price level look like Profile views and actions over time Your own photos are few or dated, and customer photos misrepresent the place Polished shoots that don’t match what arrives at the table
    Booking integration with a reservation platform Makes booking possible from the profile, and gives AI Mode something bookable The booking button on the profile; Google-sourced bookings in the reservation system The profile has no booking button, or it points to a provider you no longer use Some platforms charge per cover for bookings they class as coming through their own network
    The platform’s own listing and app Discovery inside the platform, a separate marketplace with its own ranking and, on some platforms, paid promotion Bookings with the platform’s app or network as the source Diners in your area browse on the platform Counting marketplace bookings as search bookings
    Website Brand-search conversions, occasion pages, and the text other sources and AI answers can read Analytics conversions (if the widget is tracked); organic landing pages Functions, events, set menus and groups; direct bookings without platform fees A booking widget analytics can’t see, and a menu that is only an image

    The pattern behind the table: menus and photos act on people who have already found you, the booking integration removes the step between choosing and booking, and the website carries the searches and bookings the profile can’t. Those effects compound rather than compete. A profile with great photos and no booking button sends people to your website to hunt for a widget, and some of them won’t bother.

    Measure where your covers come from before changing anything

    1. Export 12 months of bookings with their source. The source field separates your website widget, the platform’s own app or network, a Google booking button and staff entries. Check what each source value means in your system before comparing them.
    2. Count covers and revenue, not bookings. A function for 30 through the website and a table for two through an app are both one booking.
    3. Tag the profile’s links. Add UTM parameters to the website link and any booking link on the Business Profile, so analytics can separate profile clicks from other organic visits.
    4. Check whether analytics can see a completed booking. Booking widgets are often embedded from the provider’s domain or open the provider’s page, so your analytics may record the click but not the booking. Use the provider’s analytics integration if it has one, or rely on the reservation system’s source data.
    5. Read the Business Profile performance view. It shows interaction counts such as calls, direction requests and website clicks. Check which booking and menu interactions it reports for your profile before treating any of them as a trend.
    6. Give phone bookings a source. If a large share of covers arrives by phone, asking “how did you find us?” at booking is crude, but better than nothing.

    Then set a baseline. Restaurants have strong weekly and seasonal cycles (weekends, school and public holidays, local events, weather), so compare against the same weeks last year and against a channel search doesn’t touch, such as walk-ins or repeat guests.

    Hypothetical example:

    A 70-seat suburban bistro exports a year of bookings: roughly 450 a month through the platform’s own app, 300 through the Google booking button, 150 through its website widget and 100 by phone. The app bookings are marketplace bookings, not search, and some carry a per-cover fee. The 300 Google-button bookings are where profile photos, menu and reviews do their work. The website’s 150 look small until covers and revenue are added, because they include most of the group and function bookings. For this venue, “which lever?” has three answers: the profile for volume, the website for value, and the platform listing for a channel that search reporting doesn’t show.

    A test one venue can run

    Single-venue tests are noisy, so change one thing at a time and give each phase several full weekly cycles. A restaurant’s week is its natural unit of variation, so I would treat four full weeks per phase as the minimum, and avoid phases that straddle school or public holidays.

    • Hypothesis: replacing the profile menu with current text items, and adding a fresh set of dish and room photos, increases Google-sourced covers.
    • Primary metric: covers with a Google source per week, compared with the same weeks last year.
    • Secondary metrics: profile interactions, UTM-tagged website visits from the profile, website bookings.
    • Demand control: covers from walk-ins or repeat guests.
    • Confounders: a burst of new reviews, menu or price changes, fewer tables released online because of staffing, local events.

    A restaurant group can do better. Roll the same change out to venues in a staggered order and compare each changed venue with those not yet changed. That separates the effect from citywide demand far better than a before-and-after at one site.

    What would change the decision:

    • If Google-sourced bookings are a small share and most platform bookings come from the platform’s own app, the platform listing (its photos, reviews and released availability) is the bigger lever, and search is secondary.
    • If profile booking clicks are high but completed bookings are low, the problem is availability or the booking flow, not visibility.
    • If website bookings come mainly from brand searches, the website’s job is conversion (a fast page, a text menu, a widget that works on a phone) rather than content.
    • If occasion pages bring function enquiries, that is the website’s strongest case for more investment.

    Where I would spend the first few hours

    1. Make the profile bookable: confirm the booking button points to your current provider and that tables are released online for the times people search.
    2. Put the current menu in text on both the profile and the website, not only as a photo or PDF, with prices and dietary labels kept up to date.
    3. Refresh photos: signature dishes, the room during service, the entrance (so people can find the door) and the menu board if you have one.
    4. Make hours, including public holidays, match everywhere. Wrong hours send people to a closed door, and that tends to show up in reviews.
    5. Give each occasion with real demand its own page on the website (functions, set menus, events), with a clear enquiry or booking path.
    6. Set up the source tracking above and review it monthly.

    After that, let the booking data decide where the next hours go, not the channel you like best. If most covers come through the profile, invest there. If the valuable ones come through the website, invest there, even if the volume is smaller.

    Need help with this on your own site?

    Niraj Raut works with businesses in Nepal, Australia, the UK, Europe and the US on technical SEO, ecommerce SEO, local SEO and AI search.

    See SEO services

Add a reply

Replies are open to approved members. Log in or apply to join.