Duplicate guest records and how to merge them

A guest's name appears four times in the guest list. "Sarah Jones", booked online with her work email. "S Jones", a phone booking taken by the host. "Sarah J", a walk-in whose name was noted at the door. "Sarah Jones", booked again with her personal email. To the venue, she looks like four occasional guests. In fact, she is a regular who has visited eighteen times, has a shellfish allergy recorded on only one of the four profiles and is about to be sent the same "we miss you" email three times.

Duplicate guest records are one of the most common problems in any guest database, and one of the most damaging. They split a guest's history, hide their allergies, make regulars look like strangers and send the same message to the same person more than once. They also make every report about guests less accurate.

This guide covers how duplicates form, how to find them, how to decide whether two records are really the same guest, what to keep when merging and how to prevent new duplicates from forming. Every venue and figure in it is an example.

Duplicates matter because they break the guest's history into pieces. A regular who appears as four guests has four short histories instead of one long one. Nobody sees that she has been eighteen times; nobody greets her as a regular.

They also hide important information. An allergy recorded on one profile does not appear when she books under another. A preference noted by one server is invisible to the next. And campaigns go wrong: she receives the same email several times, or a "we miss you" message when she was in last week under a different record.

Duplicates form in predictable ways. A guest books online with one email and later with another. A host takes a phone booking and types the name slightly differently. A walk-in gives a first name only. A guest books under a partner's name. Two systems, such as a booking platform and the venue's own booking page, each create a record.

Each is ordinary. Over a few years, they add up to a guest database where a meaningful share of guests appear more than once.

Most duplicates follow a few patterns. The same name with different emails, usually work and personal. The same phone number with different names, often a couple who both book. The same email with a misspelled name. A name with and without a middle initial. A first name only for a walk-in, alongside a full record from a booking.

Knowing the patterns makes duplicates easier to spot, because you know what to look for.

Most guest databases have some duplicates, and a database built over several years from several booking sources may have many. Do not be alarmed by the number when you first look; be concerned only if nothing is done about them.

A useful measure is the share of profiles flagged as possible duplicates. If it falls after a clean-up and stays low, the prevention is working. If it creeps back up, look at where new duplicates are coming from: a booking source, a habit at the host stand or an import.

The aim is not a perfect database, but one accurate enough that the team trusts it and guests are recognised.

Most guest systems can flag possible duplicates automatically, by looking for records that share an email address, a phone number or a very similar name. Use that as the starting point, then check each suggested pair yourself.

Without a system to flag them, you can find many duplicates by sorting the guest list by phone number, then by email, then by surname, and looking for neighbours that look alike. The general problem of finding records that refer to the same person is known as record linkage, and the principle is the same: compare the details that identify a person, and look for matches.

Not every match is a duplicate. Two guests can share a name; a couple can share a phone number; a family can share an email. Before merging, check that the records really are the same person.

Strong signals are an exact match on email or phone number together with a similar name. Weaker signals are a similar name alone, or a shared phone number with different names. Where you are unsure, leave the records separate, or ask the guest next time they book.

Shared contact details are the most common cause of wrong merges. A couple who book with the same phone number are two guests, not one. A family email used by several people links several guests.

Merge only when you are confident the records are the same person. A wrong merge is worse than a duplicate, because it mixes two people's histories, preferences and allergies, which can be dangerous.

Couples, families and colleagues are the hardest cases. A couple who book under either name, with a shared phone number, are two guests who visit together. A family where several members book is several guests linked by one email address. Colleagues who book for each other leave records under each other's names.

Do not merge them into one profile. Keep each person as their own guest, and use notes to show how they are connected if that helps service: "Usually visits with their partner, who has a nut allergy."

Merging two people into one mixes their allergies, preferences and histories, which can be confusing at best and dangerous at worst.

When merging two records, choose which profile to keep as the main one, usually the one with the most history or the most reliable contact details. Then bring across what matters from the other: visits, bookings, orders, feedback, allergies, preferences and notes.

Where the two records disagree, such as two different phone numbers or two different birthdays, decide which is correct. If you cannot tell, keep both pieces of information in a note rather than discarding one.

When merging, check allergies and dietary needs first. If one record lists a shellfish allergy and the other does not, the merged profile must keep the allergy. Never let a merge remove an allergy.

If the two records list different allergies, keep all of them on the merged profile and confirm with the guest at their next visit. It is far better to ask a guest about an allergy they do not have than to forget one they do.

Marketing consent needs care too. If one record shows the guest agreed to receive marketing and the other does not, think about which is more recent and how each was given. When in doubt, take the more cautious view: do not send marketing unless you are sure the guest agreed.

Keep a record of the merge, so you can show later what each original record said.

When two records for the same guest disagree, the merge has to decide which is right. Two different phone numbers may both be current; two different birthdays cannot both be. Two different email addresses may be work and personal.

Keep both where both can be true, such as two phone numbers, and choose where only one can be, using the most recent and most reliable source. Where you cannot tell, note the difference and ask the guest at their next visit.

Never guess about allergies or consent. Keep every allergy, and take the cautious view on consent.

A good merge leaves a trail: when it happened, who did it, which records were combined and what was moved. That trail lets you check a merge later, explain it if a guest asks and undo a mistake if one was made.

Never delete the evidence of a merge. A guest database where records quietly disappear is one nobody can trust.

Finding and merging duplicates is part of guest data hygiene: the regular work that keeps a guest database accurate and useful. Make it a habit: an hour a month reviewing suggested duplicates, removing dead email addresses and updating out-of-date preferences.

Small, regular clean-ups are far easier than a large clean-up once a year, when duplicates have piled up and nobody remembers which guest is which.

Mistakes happen: two different guests merged because they shared a name and a phone number. When it happens, act quickly. Use the record of the merge to see what came from each original profile, and separate them again as carefully as possible.

Check allergies and preferences especially, because a wrong merge may have given one guest another's allergy, or hidden an allergy from the guest who has it. Tell the team about the correction if either guest is booked soon.

A good merge record makes a wrong merge fixable. Without one, untangling two people can be nearly impossible.

The best way to reduce duplicates is to stop them forming. When taking a booking, search for the guest first, by phone number or email, before creating a new record. When a walk-in gives their name, ask for a surname or a phone number. When a guest books with a new email, check whether they already exist.

Train the team to search before creating. A few seconds at the time of booking saves a merge later.

If you are building a guest database from bookings you already have, expect duplicates in the first import, especially if the bookings came from several sources. Clean them before you rely on the database, not after.

Our guide to building a guest database from bookings you have covers the first build.

Imports and booking platforms are common sources of duplicates. An imported list from an old system may contain guests who already exist in the new one. A platform may create a new record for each booking, even for a returning guest.

After any import, review the possible duplicates before relying on the data. For platform bookings, check whether the guest already exists when the booking arrives, and link it to their existing profile where you can.

Catching duplicates at the moment they arrive is far easier than finding them months later.

The host stand creates more duplicates than anywhere else, because it is where phone bookings and walk-ins are recorded in a hurry. Train hosts to search before creating, by phone number first, then by surname, and to ask a returning guest for the details they booked with before.

A host who spends five seconds searching saves a manager five minutes merging later, and makes sure the guest's allergies and preferences appear on the booking straight away.

Once duplicates are merged, the guest's history becomes complete: every visit, every order, every piece of feedback in one place. That is when the history becomes useful, because it shows the guest's real pattern rather than fragments.

Our guide to guest history: what the last four visits tell you covers how to read it.

A guest database full of duplicates is one the team stops trusting. Servers learn that a guest's profile may be incomplete, so they stop looking at it. Merging duplicates, and keeping them merged, is what makes guest profiles worth reading.

Our guide to guest profiles your floor team will trust covers what makes a profile reliable enough to use.

Merging duplicates pays off quickly. Regulars appear as the regulars they are, so the team recognises them. Allergies recorded on one profile appear on every booking. Campaigns reach each guest once. Reports about guests become accurate, because each guest is counted once.

The work is modest: an hour a month for most venues, more for the first clean-up. The reward is a guest database the whole team can rely on, and guests who feel known every time they visit.

A first clean-up can feel overwhelming when there are hundreds of possible duplicates. Work in small batches instead: twenty or thirty at a time, starting with the strongest matches, such as an exact email and a matching name, which can be merged quickly and safely.

Leave the uncertain cases, such as similar names with different contact details, for a separate session when you can take more care. Some will turn out to be different guests; some will need checking with the guest at their next visit.

An hour a week for a few weeks clears most of a backlog without anyone dreading it.

Decide who can merge guest records. Merging is a powerful action, because it combines histories, preferences and allergies, and a mistake can be hard to undo. Usually, a manager or a trusted senior member of staff does it, rather than everyone who can see guests.

Make sure whoever merges understands the rules: confirm it is the same person, keep every allergy, handle consent cautiously and record the merge. A short checklist beside the screen helps.

The rest of the team can help by flagging suspected duplicates when they notice them, so the person who merges has a list to work through.

Most guests never know their records were merged, and they do not need to. But sometimes a merge reveals something worth confirming with the guest: two different allergies, two different phone numbers or a preference that has changed.

Ask lightly at their next visit: "We have two numbers for you; which is best?" or "We have a note about a shellfish allergy; is that still right?" Guests are usually pleased that the venue is keeping their details accurate.

Their answer makes the merged profile more reliable, and the question itself shows the guest that the venue takes care with their information.

Merging duplicates changes your reports. The number of guests falls, because each person is counted once. The number of regulars rises, because guests who looked occasional turn out to visit often. Average visits per guest and spend per guest rise too.

Note when you did the clean-up, so the change in the figures is not mistaken for a change in trade. From then on, the figures are more accurate, and comparisons over time become meaningful.

A guest database without duplicates is kinder to guests as well as more useful to the venue. Each guest hears from you once, is greeted as the regular they are and never has to repeat an allergy because it was on a different profile. That is the real reason to keep the list clean.

An invented bistro builds its guest list from five years of bookings, taken online, by phone and through a booking platform. When it reviews the list, the system flags several hundred possible duplicates: the same email, the same phone number or very similar names.

The manager works through them over a few weeks, an hour at a time. Exact email matches are merged quickly. Shared phone numbers are checked, and couples are left as separate guests. Allergies are always kept. Each merge is recorded with who did it and when.

Afterwards, the bistro's list of regulars grows noticeably, because many guests who looked occasional turn out to be frequent. A guest with a shellfish allergy, previously recorded on only one of her three profiles, now has it on her single profile, where every booking shows it.

In servearound, a guest's profile flags possible duplicates, each with how confident the match is and why: the same email, the same phone number or a similar name. Merging asks which profile to keep, shows where the two disagree and then moves the reservations, orders and feedback onto the profile you keep, with the merge recorded on the profile, by date and by who made it.

A merged profile carries the guest's complete history, so their visits, spend, allergies, preferences and feedback appear together on every future booking. Guests can be found by name, email, phone or notes before a new profile is created. See Guest CRM.

Two or more records in the guest database for the same person.

Different emails, phone bookings with spelling differences, walk-ins with first names only, bookings under a partner's name and records from several systems.

Use the system's duplicate suggestions, or sort the list by phone, email and surname and look for matches.

An exact email or phone match with a similar name is strong; a shared phone with different names is not.

The profile with the most history, plus every visit, order, piece of feedback, allergy and preference from the other.

Yes, if two different people are merged. Merge only when you are confident.

Search for the guest before creating a new record, every time.

A short session every month keeps them under control.

Duplicate guest records split a guest's history, hide allergies, make regulars look like strangers and send the same message twice. They form in predictable ways: different emails, misspelled names, walk-ins, shared contact details and several booking sources.

Find them with the system's suggestions or by sorting the list, merge only when you are confident the records are the same person, keep every allergy, handle marketing consent cautiously, record every merge, search before creating new records and make guest data hygiene a monthly habit.