A taxonomy/model for user context with Paola Roccuzzo

A taxonomy/model for user context: defining user attributes to understand personalisation Paola Roccuzzo — Experience Design Director (Content Services), Foolproof London 2024

WIAD LONDON - MARCH 2024 A taxonomy/model for user context Defining user attributes to understand personalisation Paola Roccuzzo, experience design director for content services @Foolproof_UX

WIAD LONDON - MARCH 2024 draft classification A taxonomy/model for user context Defining user attributes to understand personalisation Paola Roccuzzo, experience design director for content services @Foolproof_UX

01 Some background

2019: GOV.UK future strategy - personalisation Our brief was to understand how a government account could remove friction for users interacting with GOV.UK. Things we had to grapple with: ● define the different outcomes of a personalised experience ● find successful examples to point to ● land a shared understanding of the term ‘personalisation’

The personalisation spectrum: 2019 ❌

In hindsight, despite using the concept in our prototypes, we didn’t fully understand digital identity. This is where I’m restarting from today.

02 The (draft) classification

Some caveats before diving in ● This is a work in progress, highly objectionable–objections welcome, this is a call for arms ● Some bits are esoteric–but technology is evolving fast ● It’s a partial view because 30 minutes is a short time–come find me for the rest if we don’t make it on time ● User context source: W3C Verifiable Credentials recommendations ● Personalisation outcomes inspo: Jeff Eaton

Facets for user context: user attribute type User attribute type Signal Credentials

Facets for user context: source Source Device System Holder

Facets for user context: source Source Device System Holder Definition The entity making the claim (so far, only human)

Facets for user context: privacy spectrum Privacy spectrum Non correlatable Correlatable via collusion Highly correlatable Data minimisation more privacy less privacy

Facets for user context: privacy spectrum Privacy spectrum Example “Is older than 18” Non correlatable Correlatable via collusion Highly correlatable Example Name, date of birth, postcode Example Government ID, shipping address

Facets for user context: level of identification Level of identification Example Returning user for analytics Returning device Authorised credentials Verified identity A system account with username, password and 2FA A system account connected to a valid proof of identity (ex. online banking account)

Enter the trust model ● A trust model implies a set of rules and standards to verify identity (and claims) which organisations agree to follow ● Not common (yet): few systems require identity verification ● When you use a digital identity system (like GOV.UK One Login) you’re entering a trust ecosystem ● In the future the use of digital identity might extend to systems that right now don’t require it

Facets for user context: trust model entities Trust model entities Holder Subject Issuer Verifier The entity that submits verifiable credentials The entity related to verifiable credential (can be ≠ holder, a parent or a carer) The entity that issues a verifiable credential for a holder The entity that verifies a verifiable credential for an issuer

Facets for user context: trust model concepts Trust model concepts Attribute that can be verified and exchanged in a trust ecosystem Verifiable credential Verifiable presentation Proof Data derived from one or more verifiable credentials The cryptographic encoding that allows to see whether a credential or a presentation has been tampered with

The personalisation spectrum: 2024 (WIP)

03 In practice

Identity verification system Recommendations

We’re now in a trust ecosystem Dynamic assembly based on verifiable credential + proof Dynamic assembly based on verifiable credential - no proof Verifiable credentials with proof

Verifiable credentials with proof Verifiable credentials needs proof

Issuer Holder (and subject) Verifier

Automated decision based on verifiable presentation

[The application journey continues on DWP]

Dynamic assembly based on verifiable presentation

Thanks