Impact Scores

Every post-release metric receives an impact score — a number from 0 to 100 that reflects how significant the change is compared to your baseline data. The higher the score, the stronger the evidence that a real change occurred. When the change moves opposite your target direction, the score is shown as a negative number (e.g., -72) so you can see both the strength and direction at a glance.

How current is the result?

Wherever VEKTIS shows a result, it also shows when the underlying data last changed — a small “Updated … ago” label (hover it for the exact time). This matters most for features VEKTIS measures automatically: if the “Updated” time is older than you’d expect, new activity may not be reaching VEKTIS, and the score you’re reading is based on stale data. For hand-entered features, it simply reflects when you last added a measurement.

Impact zones

The score falls into one of three zones:

Zone	Score range	What it means	What to do
Significant Impact	70–100	Well outside the normal range — strong evidence this is real	Act on it — very unlikely to be noise
Confirmed Impact	40–69	Outside the normal range — likely a real change	Promising — a few more data points will confirm
Signal Detected	0–39	Within or near the normal range — could be regular fluctuation	Keep measuring — not enough evidence yet

Direction coloring

The impact score is combined with direction relative to your target to determine coloring:

Scenario	Color	Meaning
Significant Impact, target direction	Green	Strong impact in the direction you wanted
Significant Impact, wrong direction	Red	Strong change, but opposite your target
Confirmed Impact, target direction	Green	Likely a real change in the right direction
Confirmed Impact, wrong direction	Red	Trending opposite your target — worth investigating
Signal Detected (either direction)	Gray	No clear signal yet
Negative Impact (wrong direction, any strength)	Red	Change is opposite your target — score shown as negative

Interpretation messages

Along with the score, VEKTIS shows how far outside the normal range the result falls:

Range	Interpretation
Within normal range	“Looks like normal fluctuation”
Slightly outside	“Something might be happening”
Clearly outside	“This looks real”
Far outside	“Strong signal — this isn’t random”

When you can’t calculate impact

Impact can’t be calculated when:

• Fewer than 2 baseline data points — VEKTIS needs at least 2 measurements to establish a normal range
• No variation in baseline — If all baseline values are identical, there’s no “normal range” to compare against

Tip

If you see “Need more baseline data,” add more baseline metrics to your dev item. Four or more data points give the most reliable results.

Impact per customer

The score above is the overall picture. For features that collect per-customer data, the same scoring is also applied to each customer individually — so you can see whether an impact is broad or concentrated in a few accounts. That view lives in the Customer Breakdown on the dev item’s detail page. See Customer Breakdown for how to read it.

Tip

In the per-customer table, the three zones are shortened to Significant, Confirmed, and Signal to fit the row — they mean the same thing as the full labels above.

Practical guidance

You see	What to do
Significant Impact (green)	Celebrate — strong evidence your feature worked as intended
Significant Impact (red)	Investigate — something changed significantly, but not in the direction you wanted
Confirmed Impact (green)	Promising — keep measuring to see if the pattern holds
Confirmed Impact (red)	Worth attention — keep measuring, and consider investigating if the trend continues
Signal Detected	Keep measuring — one data point isn’t enough. Add more post-release measurements over the coming weeks