Colt Replaced Transactional NPS With Net Easy Score

At a glance

• Colt was running transactional NPS across a network spanning roughly 40 countries, and finding it did not predict what customers went on to do.
• The programme moved to Net Easy Score, which measures how hard the customer had to work rather than how they felt afterwards.
• NES reached 71.4 in Europe and 46 in Asia, with Service Management improving 17 points over 24 months.
• Surveys were cut to three elements: the effort question, a comment box, and drilldowns only where the answer justified them.

The situation

Colt is a digital infrastructure business operating across roughly 40 countries, with a network reaching around 32,000 buildings in about 230 cities. Its customers are enterprises, and its relationships are managed, contractual and long-running.

The company was measuring transactional NPS after service interactions. The scores arrived reliably. They just did not tell anyone very much.

The specific failure was predictive. Accounts that scored well were not reliably renewing or expanding, and accounts that scored badly were not reliably leaving. A metric that does not correlate with behaviour is an expensive way to generate a number for a slide.

Asking an enterprise customer how likely they are to recommend you, immediately after a support ticket, measures their mood. It does not measure whether the relationship is working.

Why transactional NPS was the wrong instrument

NPS asks about advocacy, which is a relationship-level sentiment. Applying it to a single transaction asks a customer to summarise a whole relationship on the basis of one interaction, and the answer moves with whatever just happened.

In a B2B setting this gets worse. The person answering is often not the person who decides on renewal. The account has a commercial reality, and a support engineer's ticket experience is one small input to it.

What actually predicted behaviour was effort. How much work the customer had to do to get something resolved, and whether that work felt proportionate.

What we changed

The programme moved to Net Easy Score, calculated as the percentage of customers reporting the interaction as easy minus the percentage reporting it as difficult. Not an average of a satisfaction scale, which flattens exactly the extremes that predict behaviour.

The survey itself was reduced to three elements:

• The effort question
• An open comment box
• Follow-up drilldowns triggered only where the score or the comment warranted it

Everything else came out. In an enterprise setting, survey length is not a neutral cost. Every additional question is one more reason for a senior stakeholder to close the tab.

Feedback was then routed into the alerting and workflow layer so that a difficult interaction generated an action rather than a data point.

The result

Net Easy Score reached 71.4 in Europe and 46 in Asia. Service Management improved by 17 points over 24 months.

The regional spread is worth reading rather than averaging. A 25 point difference between two parts of the same business, measured the same way at the same time, is a diagnostic finding in itself. It identifies where the operational friction actually sits, which a single global number would have hidden completely.

Methodology: the figures come from a programme that ran over 24 months. The sample size is held under client NDA.

Effort predicts behaviour. Satisfaction predicts how someone felt an hour ago. For a contractual B2B relationship, only one of those is worth measuring after a service call.

Amitayu Basu, CEO and Co-founder, Numr

Why Net Easy Score, specifically

Numr scores effort using the net method rather than a mean, and the distinction matters more than it sounds.

Averaging a seven-point effort scale produces a comfortable middling number that moves slowly and hides its own extremes. The customers who found you very difficult are the ones about to act, and an average is designed to bury them among the people who found you fine.

The net calculation keeps both ends visible. A score of 71.4 means promoters of ease substantially outnumber its critics. It is a different claim from an average of 5.2 out of 7, and a considerably more useful one.

What transfers to other enterprise programmes

Match the metric to the level of the question. Relationship-level metrics for relationship-level decisions, effort metrics for transactions. Asking a transactional NPS question is not a small methodological preference, it is measuring the wrong thing.

Read the regional spread before the headline. The gap between Europe and Asia was more actionable than either number on its own.

Shorten the survey before you try to raise the response rate. In enterprise accounts, length is the constraint.