You calculate a clean average, say it out loud, and move on. The interviewer already knows the blend can be lying. A 14% margin, a 6% growth rate, or a $40 average order can be two opposite businesses sitting in one number.

Segment before you average. Split the mix, see which piece is carrying the blend, and only then decide. If you calculate the next step off the blend, you may solve the wrong problem.

Why the blend hides the call

• Opposite segments cancel. One line can be growing while another is shrinking, and the average looks flat.
• Weight matters as much as the rate. A high margin on a small slice does not make the whole business healthy.
• The action changes. A price move, a cut, or an exit aimed at "the average customer" often hits the wrong segment.

The sequence

1) Pause on the average

Margin, growth, price, utilization, conversion. If the number is a blend, do not treat it as the answer yet. Say that you want the mix before you decide.

2) Name one cut

Pick the split most likely to hide opposite stories: product line, region, customer type, channel, or store format. One cut. Not five.

3) Calculate level and weight

For each piece, calculate the rate and how much of the total it represents. A 28% margin on one third of units is a different business from a 28% margin on nearly all of them.

4) Say which piece drives the decision

Name the segment that is carrying the average, and what you would do differently because of the split. Then, and only then, calculate the next step.

Want reps on this? Practice segment-before-you-average drills with CaseCrunchAI and get feedback on whether you decided off the blend or off the split.

A 30-second script

Prompt math: blended gross margin is 14%.

"Blended gross margin is 14%. I am not calling that healthy yet. Premium is about a third of units at a 28% margin, and the value line is about two thirds at 7%. The blend is being carried by premium. So the decision is not an across the board price increase. It is whether the value line is worth keeping."

Same total. Different call. The split did the work.

What to skip

• Averaging the averages. If the segments are different sizes, a simple mean of the two rates is a second wrong number.
• Five cuts at once. Product, region, and channel in the same breath is a stall. Pick the cut that would flip the recommendation.
• A split with no decision. If both segments point to the same action, say that in one line and move. Segmenting is for the call, not for extra math.

10-minute drill

1. Pick three blended numbers from old cases: a margin, a growth rate, and a price or utilization figure.
2. For each, set a 30-second timer. Out loud, name one cut and sketch two segments that pull in opposite directions, with a rough weight for each.
3. State which segment would flip the recommendation if it were most of the mix.
4. Repeat the one you rushed until the split lands before any next calculation.

A blended number is a hiding place, not an answer. Split it, then decide.

Start practicing free at CaseCrunchAI and run timed calculate drills that force the segment before the average.