TapCub is live — analytics, insights and live chat, free on one platform
Glossary

What is a single customer view? Data layers, ID rules and a merge example

A single customer view is one consistent profile and timeline per person. The five kinds of data it needs, why a stable ID comes before the view, and a merge example from 1,200 anonymous IDs to 1,020.

Chen Yu Reviewed by Emma Collins Also called single view of customer Last updated Nov 17, 2026 Glossary

Quick answer

A single customer view is one consistent profile and timeline for each person, combining identity, behaviour, transactions, conversations and attributes from every system. It exists only when the same person has one stable ID; joining tables without that ID does not create it.

Formula: ID reduction = 1 − (people + unmatched IDs) ÷ anonymous IDs; identified share = identified IDs ÷ all IDs

What is a single customer view?

A single customer view (SCV, also called a single view of the customer, or a customer 360 view) is one consistent record of everything an organisation knows about one person, available in one place. Instead of finding a person in the web analytics tool under one ID, in the order system under an email and in the support inbox under a name, you open one profile and see all of it on one timeline.

The idea has a regulatory root as well as a marketing one. In the UK, deposit-taking firms must be able to produce a single customer view file with the information the compensation scheme needs to pay customers if a firm fails; the deadline for providing it is 24 hours. Marketing and support teams borrowed the term for the same reason: to stop treating one person as several strangers.

The core of this entry: a single customer view is not what you get by joining all your tables into one. It is what you get when the same person has one stable ID, so behaviour, orders and conversations hang on the same timeline. Build the ID first, then the view. A warehouse full of unjoined tables is a pile of data. One person ID turns it into a view.

The view sits next to three related terms, and it helps to keep them apart:

TermWhat it isHow it relates to the view
Single customer viewThe result: one profile and timeline per personWhat a salesperson or support agent looks at
Identity resolutionThe method that links records and IDs to the same personHow the view is assembled
Identity graphThe data structure that stores which IDs belong togetherThe lookup the view relies on
Customer data platformA product category that collects data, resolves identities and exposes profilesOne way to build and serve the view

This page is about the view itself. The linking algorithms are covered in the identity pages.

The five kinds of data in the view

A useful view has five layers. A profile with only one or two is a contact record, not a view.

LayerWhat it holdsTypical examplesWhere it usually comes fromJoin key
IdentityWho the person is across devices and systemsLogin ID, email, phone, device IDsSign-in events, account systemLogin ID
BehaviourWhat they didPages viewed, events, sessions, features usedWeb, app and server analyticsDevice ID until login, then login ID
TransactionsWhat they bought or paidOrders, refunds, subscriptionsOrder or billing system, purchase eventsCustomer or login ID
CommunicationWhat was said to and by themChat conversations, support tickets, emailsLive chat, help desk, email toolVisitor ID, email or login ID
AttributesWhat is known about themName, plan, country, tags, segmentsProfile fields, rules, importsLogin ID

Notice that every layer needs a key that matches the identity layer. The most common reason a "customer 360" fails is not missing data but a layer that cannot be joined: purchases stored under an email, behaviour stored under an anonymous device ID, and no record linking the two.

How the ID connects the layers

The rules for linking are simple to state, and strict:

  1. A login ID beats everything. When a person signs in, you know that this device, this session and this account are one person. That is a deterministic link: it is either true or it is not. See user ID and identify for how the call works.
  2. Before login, a person is a device. An anonymous visitor ID describes a browser or app install, not a person. The same person on a laptop, a phone and a tablet is three IDs until something ties them together.
  3. Merge on login, not on guesses. Link devices when the same login ID appears on each. Matching on similar names, IP addresses or screen sizes (probabilistic matching) produces false merges, in which two people are fused into one profile, and that is harder to repair than a missed merge.
  4. Anonymous visitors do not get a profile. If someone never signs in, they stay in aggregate counts. This is a feature: you cannot invent a view of someone who never identified themselves.
  5. Backfill, then keep going. When a device first logs in, its earlier anonymous activity can be attached to that person, so the timeline starts at the first visit, not at the login.

The honest consequence: the view covers only people who identified themselves. If 5% of your visitors log in, you have a view of 5% of your audience and aggregates for the rest.

Worked example: five anonymous IDs, one person

Sample data: over one week, a site's analytics counts 1,200 anonymous visitor IDs. 300 of these IDs belong to people who signed in at some point; the other 900 never did. The 300 IDs turn out to be 120 people, because many people used more than one browser or device (300 ÷ 120 = 2.5 IDs per person).

One of those people, as a concrete case:

Anonymous IDDeviceFirst seenBefore mergeAfter merge
a1Laptop, ChromeMonVisitor 1Person P
a2Phone browserTueVisitor 2Person P
a3Laptop, Chrome after clearing storageWedVisitor 3Person P
a4Phone appThuVisitor 4Person P
a5TabletFriVisitor 5Person P

Five IDs on three devices become one person once login connects them. Now the whole week:

CountBefore mergeAfter mergeCalculation
Visitors (people or unmatched IDs)1,2001,020120 people + 900 unmatched IDs = 1,020
Reduction–15%1 − 1,020 ÷ 1,200 = 15%
Share of IDs that are identified–25%300 ÷ 1,200 = 25%
Share of the merged count that is identified–11.8%120 ÷ 1,020 = 11.8%
Conversion rate, 48 buyers (all signed in)48 ÷ 1,200 = 4.0%48 ÷ 1,020 = 4.7%The denominator shrank, not the buyers

Two lessons. First, merging changes the visitor count by 15% here, so every rate that divides by visitors changes too: conversion looks 0.7 points better without a single extra sale. Second, only 120 of the 1,020 (11.8%) are people you have a real view of; the other 900 are still IDs. If you want a bigger view, the lever is getting more visitors to sign in at a sensible moment, not a cleverer matching algorithm.

What goes wrong, and how to check

  • One row per person is not a view. A table with one line per person (name, email, last order) is useful, but a view needs the timeline underneath: what happened, in what order, on which device.
  • Probabilistic merges you cannot undo. If two people are fused by a guess, support will quote one person's order to another. Prefer deterministic links, and keep a log of every merge and split so a wrong one can be reversed.
  • Shared devices. A family laptop or a shop terminal has many logins on one device. A good rule: when a different login appears on a device, keep the earlier activity with the earlier person and start the new person's timeline from that point.
  • Stale views. A view that updates nightly is wrong during the support call. Decide what has to be live (conversation context, last order status) and what can wait.
  • Ignoring deletion. A view is personal data. Under the EU's GDPR, a person can ask for erasure of their personal data under Article 17, and that request has to reach every layer of the view, not only the one table you remembered.
  • Counting the unknown as known. A report that says "customers" but includes unmatched visitor IDs overstates the base. Label identified people separately.

A quick check before you call something a single customer view: pick five real customers and, for each, try to answer "what did this person do before they bought, and what have we said to them since?" If you cannot answer in under a minute, you have a set of tables, not a view.

Building the view in TapCub

Take support's question: "who am I talking to, and what have they done on the site?" Here is how the pieces line up in TapCub. TapCub is not a CRM: the view covers what TapCub collects (behaviour, events you send, chat), not contact records or invoices kept elsewhere.

  1. Turn on identification and call login. Profiles need identification, and whether a login is bound to a profile depends on the settings in the Privacy center. Then call tapcub.login('user-id', { $name: '…' }) on a successful sign-in. See identify users.
  2. Open the person in All-platform analytics › Users › People. Searching by name or login ID finds them. The profile shows their devices and login IDs, segments, tags and accounts on the left, and a day-by-day timeline on the right that merges all bound devices. Activity from before they logged in is labelled as an anonymous segment, so the timeline does not start at login.
  3. Check how much merging happened in All-platform analytics › Users › Identity map. A difference card compares visitor, device and person counts, in the same spirit as the 1,200 versus 1,020 above, and the mapping log records every bind, backfill, unbind and split so a wrong link can be reversed by an owner or admin.
  4. Give support the same context. Chat and analytics use the same visitor ID, so the Live chat › Inbox side panel shows the visitor's current page, referrer, analytics trail from the last two days and past conversations, with a View in analytics button to the visitor profile.
  5. Know the limits. Anonymous visitors do not get a profile or enter segments. On the strict privacy preset the visitor ID rotates daily, so the two-day trail breaks at midnight; the balanced preset keeps it.

See also: What are data silos? Causes, costs and how to connect them.

See also: Customer Relationship Management Strategy: Decide the Data Before the Tool.

See also: What is data unification? Steps and a before/after example.

Sources

Was this explanation clear?

We use your answers to fix and expand this page.

Chen Yu

Technical Engineer

Focuses on tracking code, SDKs, and API integrations, with practical guidance on implementation checks and troubleshooting.

FAQ

Questions people ask

Short answers to the follow-up questions that come up most.

Still have a question?

Chat with the team behind TapCub — we usually reply within a few hours on business days.

What is the difference between a single customer view and customer 360?

In practice the terms are used for the same goal: one profile per person. Customer 360 stresses the breadth of data in the profile; single customer view stresses that there is one consistent record. Either way it needs a stable person ID to join the data.

How do you build a single customer view?

Give each person a stable ID first, normally a login ID. Then link each data layer (behaviour, transactions, conversations, attributes) to that ID, merge devices when the same login appears on them, and show the result as one timeline per person.

Can anonymous visitors have a single customer view?

No. An anonymous ID describes a browser or app install, not a person. Until someone signs in, their activity stays in aggregate counts. On first login, earlier activity on that device can be attached to the person.

Is a customer data platform the same as a single customer view?

No. A customer data platform is a product that collects data and resolves identities; the single customer view is the profile it produces. You can also build the view from your own database if every layer joins on the same person ID.

See this number on your own site

Add one script tag for web and apps, and the first reports fill in within minutes.

See your data clearly. Find your growth.

Every click, backed by data. Install one line of code and see your first numbers in a minute.

No credit card · Free plans for analytics and chat · Cookieless analytics