Intelligence

Which change actually improved your numbers?

Turn menu, visit, and service records into comparisons your team can use to make a decision.

Osoul manager dashboard on a desktop screen

Where the opportunity gets lost

A busier Saturday does not prove a new offer worked. Changes in covers, discounts, or party size can make a headline number misleading.

How Osoul supports the work

Osoul brings performance records into a common view. Define the metric, choose comparable services, and review the result alongside changes in demand and service conditions.

Put it into practice

Dish interest compared with recorded orders
Revenue per cover compared across equivalent services
Return visits grouped by first-visit period

Pick one change and one primary KPI. Keep the comparison period and calculation consistent, then review the trade-offs.

Know what to measure

Revenue per cover, dish contribution, and relative change against a baseline.

Revenue per cover

Net sales ÷ covers served

Check whether growth comes from more guests, more spending, or both.

Relative KPI change

(New value − baseline value) ÷ baseline value × 100, for a nonzero baseline

Separate relative percentages from percentage-point differences.

Make the denominator visible.

Worked example, not a customer result. Figures below are hypothetical; the source explains why controlled comparisons matter.

10% → 12%

A 2 percentage-point increase is a 20% relative lift

Illustration: 100 orders from 1,000 menu sessions versus 120 from 1,000. A fair comparison also accounts for differences in guests, timing, and availability.

For your venue: define the conversion event and compare equivalent groups.

Microsoft Research · Controlled experimentation →