Relational survey

Customers form an overall view of a brand from many interactions, plus things no single interaction captures: price, reputation, product range, how problems were handled months ago. A relational survey asks about that overall view. It is sent on a schedule to a defined group of customers — all active customers, a key segment, a random sample — rather than being triggered by an event.

Its counterpart is the transactional survey, sent right after a specific interaction. The two answer different questions, and mature programs run both:

	Relational survey	Transactional survey
Question	How do you feel about us overall?	How did this interaction go?
Trigger	Calendar (quarterly, twice a year)	An event (visit, purchase, call)
Typical metrics	NPS, overall satisfaction, attribute ratings	CSAT, CES, transactional NPS
Length	A few minutes	Under a minute
Best for	Strategy, trend, loyalty drivers	Operations, service recovery

Relational NPS asks the recommendation question about the company as a whole — “How likely are you to recommend [company] to a friend or colleague?” — at a fixed rhythm. NPS = % promoters (9–10) − % detractors (0–6). Because it is measured on the same population at regular intervals, it is the number most companies use to report loyalty to leadership.

Example: a bank runs relational NPS every six months by email to all retail customers. In wave one, 1,800 customers answer: 40% promoters and 25% detractors, NPS 15. In wave two, after a branch-hours change and a new complaint process, 1,900 answer: 44% promoters and 21% detractors, NPS 23. The attribute questions show the improvement came mainly from “problem handling” — consistent with what the transactional surveys had already shown.

• Loyalty: “How likely are you to recommend us to a friend or colleague?” (0–10)
• Reason: “What is the main reason for your score?” (open text)
• Overall satisfaction: “Overall, how satisfied are you with [company]?” (1–5)
• Attributes: ratings for value for money, service quality, ease of doing business, problem handling, digital channels.
• Intent: “How likely are you to still be a customer in 12 months?”
• Improvement: “What one thing should we improve?”

Keep it to a few minutes. Attribute ratings are what make the survey useful: correlating them with the NPS answer (key driver analysis) shows which parts of the experience move loyalty most.

Two to four times a year is typical. More often invites survey fatigue and gives changes too little time to show; less often means problems are found late. Whatever the rhythm, keep these constant between waves so the trend is real:

• the same core questions and scales,
• the same channel (email results and kiosk results are not directly comparable),
• the same population definition, and
• exclusion of customers surveyed very recently by a transactional survey.

In Qmeter, relational surveys are typically sent as email or SMS campaigns to a customer list or segment, with opens, clicks and responses tracked per campaign; results sit alongside transactional feedback in the same reports. See survey campaigns.

Relational scores are most useful when read next to transactional data from the same period. Four combinations are common, and each points to a different action:

Relational score	Transactional scores	Likely meaning
Rising	Rising	Service improvements are reaching customers and changing their overall view
Stable or rising	Falling	Early warning: interactions are getting worse, the relationship has not caught up yet
Falling	Stable	The cause is probably outside individual interactions — price, product, reputation, competition
Falling	Falling	A service problem that has already reached overall loyalty — act on the transactional drivers first

This reading only works if both surveys tag responses with the same branch, segment and period, so they can be compared on one report.

• Using it as the only survey. A twice-yearly score cannot tell you which branch had a bad Tuesday.
• Changing questions between waves, which breaks the trend.
• Not closing the loop. Relational detractors deserve follow-up too, even though the issue may be months old.
• Reading the score without the drivers. NPS moved — but because of price, service or product?