The Do Feel Connect lens.

Chapter 1 broke the single-score habit. This chapter replaces it with the three-dimensional lens the whole system runs on, and gives every team the same language.

The three dimensions

• Do (behaviour): what customers actually do. Visit frequency, spend patterns, channel choice, engagement.
• Feel (sentiment): how customers feel about doing it. Feedback, reviews, complaints, resolution satisfaction.
• Connect (relationship strength): how bonded they are. Trust, shared preferences, advocacy, resilience after failure.

Each dimension has a formal name (Behaviour, Sentiment, Relationship Strength); Do, Feel, Connect is the working shorthand every team uses.

The power is in the combinations

No dimension means anything alone:

• High Do, low Feel: convenience, not loyalty. A churn risk hiding inside your best-customer report.
• High Feel, low Do: untapped intent. Demand blocked by a barrier, not a relationship problem.
• High Do, high Feel, high Connect: durable loyalty. Protect it, deepen it, never take it for granted.

Trust Equity

Trust Equity is the accumulated trust a customer holds in the brand, built through consistent delivery, recognition, and well-handled failures. It is the asset the whole framework protects: high Trust Equity buys resilience to failure, faster recovery, and advocacy. It is also spendable, and a generic voucher sent in response to a specific complaint burns through equity that took years to build. Connect is where Trust Equity shows up in the data.

What each dimension is not

Most implementation mistakes are category errors, so draw the boundaries early:

• Do is not value: a high-spend customer with lengthening visit intervals is a decline signal wearing a good disguise. Do reads patterns, not totals.
• Feel is not NPS alone: one survey score is a single, lagging, response-biased signal. Feel blends satisfaction, review sentiment, complaint rate and severity, and resolution experience.
• Connect is not tenure or consent: ten years on the database, or a ticked marketing box, proves nothing. Connect must be observed: a return after a failure, a referral made, a preference volunteered and used.

From combinations to states

Each dimension is banded Low, Medium, or High. The eight named states come from the High and Low patterns, and every one is a different relationship requiring different treatment. Medium bands are not forced into a state: they raise the transition flags chapter 3 describes, so ambiguity is managed rather than hidden. That is the whole architecture: chapter 3 names the eight, chapter 4 makes the banding honest, and chapter 5 attaches the playbooks.

Applying the lens

Map your current data estate against the three dimensions. Almost every operator finds the same shape: rich Do data, patchy Feel data, and almost no deliberate Connect data. That blind dimension is not a reason to delay; it is the first finding, and chapter 4 shows how confidence scoring handles it honestly.

What good looks like

Every loyalty conversation and every report shows three scores, not one. Teams describe customers in state language ("high Do, low Feel") instead of value language ("a good customer").

Chapter checklist

• Map your data sources to Do, Feel, and Connect.
• Name your blind dimension and its biggest data gap.
• Identify one recent decision that a three-dimension view would have changed.