One number to rule them all - Your own CX index

A mature survey program usually ends up with three numbers. NPS from the quarterly relationship survey. CSAT from the follow-up after every support ticket. CES from onboarding. Each answers a different question — would they recommend us, were they satisfied, was it easy — and each is useful on its own.

Then someone senior asks how customers are doing. Not three numbers on three scales. One.

The tempting answer is to average what you already have. Resist it. NPS runs from -100 to +100, CSAT is a percentage, CES is an average rating — and NPS isn’t really an experience measure at all. It’s a loyalty outcome, the thing your experience is supposed to produce. Keep it out of the index; you’ll want it later as the yardstick.

The formula

Every serious CX framework converges on the same three dimensions: did it work, was it easy, and how did it feel? Forrester’s CX Index measures exactly these — they call them effectiveness, ease, and emotion — but its component weights are proprietary and industry-specific, so you can’t recreate it. The Temkin Experience Ratings turned the same three dimensions into an open standard: a published formula that ranked hundreds of US companies for a decade, and one you can copy exactly.

Ask three questions about a recent interaction, each on a 1-7 scale:

1. Success — To what extent were you able to accomplish what you wanted to do?
2. Ease — How easy was it?
3. Emotion — How did you feel about the experience?

Score each question as a net: the share of 6-7 answers minus the share of 1-3 answers. Then average the three nets:

Experience Index = (Success + Ease + Emotion) ÷ 3

where each component = % answered 6-7 - % answered 1-3

Say 65% rate success 6-7 and 5% rate it 1-3: Success nets +60. Ease nets +50. Emotion nets +35. Your index is (60 + 50 + 35) ÷ 3 = 48.

The published ratings give that number meaning: 80 and above is excellent, 70-79 good, 60-69 okay, and below 60 poor. A score of 48 is firmly in the red, and the component scores already tell you why: emotion is dragging it down, which is exactly what the research keeps finding. In both Temkin’s and Forrester’s data, emotion is the strongest driver of loyalty in nearly every industry, and the component companies invest in the least.

This is that same example on the dashboard — the components are live, and you can click through each tile to see the details:

The weights are yours

Equal weights are the right default. They’re the published standard, they need no justification in a board deck, and they keep your number comparable over time. But they’re a default, not a law — the index should reflect what your company actually values.

If emotion is what drives repeat business in your market — Forrester’s research says it usually is — weight it double: (Success + Ease + 2 × Emotion) ÷ 4. If you sell to procurement teams where “it worked and it was fast” wins renewals, weight success more heavily instead.

There is one honest way to settle the argument: pick your weights, then compare the index against the outcome you actually care about — NPS, renewal rate, or repeat purchase. The weighting that best tracks your outcome is the right one for you. That’s also the proper job for NPS here: the yardstick the index is validated against, not an ingredient inside it.

Build it in five minutes

This index is exactly why we built combined KPIs, which are now out of preview and included in every workspace. The recipe maps one-to-one:

• Create three metrics in Settings → Metrics using Create your own KPI: Success, Ease, and Emotion. Each is a 1-7 scale scored as Net score (% good - % bad) — mark 6-7 as good and 1-3 as bad. That reproduces the published methodology.
• Link each metric to its question in the survey editor. The three questions can live in one survey or several; the metrics collect responses across all of them.
• Add a Blended score named Experience Index with the three metrics as inputs, each weighted equally — or apply your own weighting. Set the health bands at 70 and 60 to match the published ratings.

From the first response, the index becomes a live tile with its own trend chart, with the three component metrics only one click away. No exports, no spreadsheets, no monthly ritual.

Two more metrics that usually live in spreadsheets

An expectation gap. Service-quality researchers have measured quality as perception minus expectation since the SERVQUAL instrument in the 1980s: ask what customers expected, ask what they got, subtract. That’s two rating metrics and one Difference combined metric. A negative number means you’re under-delivering, and the health bands make it immediately obvious.

A floor. Averages hide outliers — a strong region can paper over a failing one. A Worst of metric tracks the lowest of several inputs, so the tile shows your weakest market’s score instead of the comfortable mean. It turns red the day any one region slips, which is exactly the day you want to know.

Where to start

Open Settings → Metrics and press Add Metric; the combined options are at the bottom of the dialog. If you’re starting from zero, the standard NPS, CSAT, and CES presets and the NPS template are all one click away. The KPIs feature page covers the whole system, including benchmarks, custom scoring, and health bands.