Tech, ExplainedEnterprise SaaS and software
Concept

Identity resolution

Many traces, one person.

HOW IT WORKSThe company collectsidentifiers: email,phone, deviceID and cookieSoftware comparesthem, often ashashed valuesExact matchesare joined:deterministic matchingRecords withweaker sharedsignals are joinedwith a probabilityAll the linkedidentifiersform one profileOne person stops looking like several strangers, soevery message and offer reaches the right profile.HOW IT WORKSThe companycollectsidentifiers:email, phone,device IDand cookieSoftwarecomparesthem,often ashashed valuesExact matchesare joined:deterministicmatchingRecords withweaker sharedsignals arejoined witha probabilityAll thelinkedidentifiersformone profileOne person stops looking like several strangers, soevery message and offer reaches the right profile.HOW IT WORKSThe company collects identifiers: email,phone, device ID and cookie↓Software compares them,often as hashed values↓Exact matches are joined:deterministic matching↓Records with weaker shared signals arejoined with a probability↓All the linked identifiersform one profileOne person stops looking like several strangers, so everymessage and offer reaches the right profile.
An email address, a phone number, a device identifier and a cookie are compared, exact and likely matches are joined, and the result is one profile for one person.
opens a page inside this one a page in the same scene a page in another scene a concept page a label only, with no page

Identity resolution is joining the scattered records of one person, such as an email address, a phone number, a device identifier and a cookie, into one profile. A company often sees the same customer as several strangers: one who signed up on the website, one who paid at a store and one who uses the app. Identity resolution works out that they are the same person.

Here is how it works in general. The software collects identifiers from every place where the company meets customers. An email address comes from a sign-up form. A phone number comes from the billing counter. A device identifier, a code that marks one phone inside an app, comes from the app. A cookie, a small file that the browser stores, comes from the website. The software then looks for links between them. When the same phone number appears in a store purchase and in an app login, the two records are joined. This is called deterministic matching, because the link is certain. When two records share only weaker signals, such as the same home internet connection and similar browsing times, the software may join them with a probability. This is called probabilistic matching, and it can be wrong. All the identifiers that are linked to one person together make up an identity graph.

To protect privacy, companies often compare hashed identifiers. Hashing turns an email address into a fixed string of characters that cannot be turned back into the email. But the same email always produces the same string, so two companies can find a match without sending the email itself. Hashing is not complete protection, though. Someone with a list of emails or phone numbers can hash each one and look for the same string.

Take, for example, a jewellery retailer with stores and a website. Its customer data platform learns that a store customer and an online visitor are the same woman, so it stops sending her a welcome offer that she has already used.

Consider a shopper on a small brand's website. One-click checkout recognises her phone number from her purchases on other shops, and fills in her address in one step.

Say an advertiser wants to reach its own customers on news websites. An identity graph links each customer to their devices, and a demand-side platform bids to show them ads.

How is identity resolution different from identity verification? Verification proves that a person is who they claim to be, for example by checking a PAN card and a selfie when an account is opened. Resolution proves nothing about the person. It only decides which records belong together. The records it joins are usually the company's own first-party data.

Appears in

Enterprise SaaS and software
E-commerce, travel and logistics
Gaming, streaming and media
Tech, Explained is a part of ingrid.fyi. Every page explains one piece of technology in plain words.