Daniel Christopher
01
BCU Digital Marketing ProjectsFlock Together, Nando’s, Mailchimp, Peaky Blinders and degree modules
Academic projects02
Selected workDigital builds, independent projects and MSc case studies
Page03
Credit Risk AnalyticsLoan affordability regression on 44,986 records
Team case04
Ecommerce Business IntelligencePower BI, Power Query and DAX on Olist data
Team case05
Decision IntelligenceDashboard, scenarios and optimisation
Team case06
Customer IntelligenceAI personalisation, trust and loyalty
Individual case07
Data ManagementDFD, ERD and SQL prototype
Team case08
Predictive AnalyticsChurn model comparison and validation
Individual case09
This portfolio websiteIndependent Next.js and TypeScript digital build
Digital project10
How I work with AIClaude Opus 5.5 leads, Codex codes, Claude helpers research, review and write
Page11
Restaurant service mixIndependent public data Python analysis and reproducible code
Independent project12
Olist dashboardInteractive order, delivery and category reporting from the cleaned workbook
Interactive analysis13
MSc dissertationAI personalisation, trust and loyalty: method and results
Page14
ProfileBilingual background, skills and credentials
Page15
Digital MarketingCustomer context, measurement and strategy
Profile section16
Role fitSix evidence linked career directions
Profile section17
ContactEmail, LinkedIn and location
Page
Search verified workEsc to close
Analytics, customer research and digital builds, with methods and ownership stated clearly.Manchester, UK · 2026
MSc dissertation · individual research · submitted August 2026

Personalisation and loyalty.
Where does trust fit?

My individual MSc dissertation examined the relationship using 139 eligible survey responses. Trust remained material in the joint loyalty model. The findings show associations, not proof of causation.

The dissertation question

Does trust statistically mediate the association between perceived AI driven personalisation and customer loyalty?

Confirmed individual research using 139 complete, eligible UK online shopper responses collected through Qualtrics on 7 and 8 August 2026.

Analytical workflow

The six stages used to complete the dissertation analysis.

  1. 01

    Question

    Define the relationship and the proposed mediating role of trust.

  2. 02

    Collect

    Qualtrics survey of eligible UK online shoppers with consent and ethics controls.

  3. 03

    Prepare

    Eligibility, completeness, duplicate, range and labelling checks across 45 variables.

  4. 04

    Validate

    Reliability, KMO, Bartlett’s test and exploratory factor analysis.

  5. 05

    Model

    Correlation, HC3 robust regression and 10,000 sample bootstrap mediation.

  6. 06

    Interpret

    Translate associations into trust centred recommendations and visible limits.

Selected results

Trust remained material in the joint loyalty model.

The pattern was consistent with an indirect, trust mediated association, although the cross sectional design does not establish causation.

TestResult
Personalisation and trust correlationr = .591 · p < .001
Trust model · personalisation coefficientR² = .349 · B = .575
Joint loyalty model · trust coefficientB = .526 · p < .001
Joint loyalty model · personalisation direct coefficientB = .048 · p = .552
Indirect effect through trust.303 · 95% CI [.199, .422]
01

Data quality

Consent, eligibility, completeness, range and duplicate checks; raw export preserved; locked SPSS dataset and cleaned CSV cross checked.

02

Reproducibility

Structured variable labels, reproducible SPSS syntax and Python used only for two documented figures.

03

Privacy

Raw participant data is not public. The portfolio reports aggregate methods and findings without exposing individual responses.

04

Claim discipline

Cross sectional convenience and snowball sampling supports association, not causation or population wide representation.

Decision implication

Trust should be included when personalisation performance is assessed.

Explain personalisation, provide meaningful controls, minimise data use, strengthen security and monitor trust alongside verified behaviour. Test changes through staged experiments before treating the relationship as causal.

View the business caseRead the one page research brief ↗