StudyResearchStatistics

Learn Statistics With a Story

The online shop that averaged 30 minutes and still missed its courier

You run fulfilment for an online gift shop. The courier arrives 37 minutes after packing starts. Your dashboard says the average packing time is 30 minutes. Can you tell the owner that every parcel will be ready?

Follow the same packing records through each chapter as the shop asks a new question.

1. One average, three packing lines

The owner asks you to plan labour for tomorrow. Lines A, B and C each handled ten parcels and logged 300 minutes of packing time. You divide 300 by ten. At that rate, 20 similar parcels would require about 600 minutes of packing time in total.

Mean = total packing time ÷ number of parcels = 300 ÷ 10 = 30 minutes

“Good,” the owner says. “So a parcel usually takes thirty minutes?”

Check the packing records before you answer
LinePacking times in minutesMeanWhat the data tells you
AAll ten took 30 minutes30The mean describes every parcel on this line.
BNine took 20 minutes, one took 12030The mean accounts for the long order when you plan labour. Most parcels finished in 20 minutes.
CFive took 25 minutes, five took 3530There are two equally common packing times. No parcel took 30 minutes.

Why bring in the median?

Customer support asks what a typical order on Line B experienced. Sort its ten times. The fifth and sixth are both 20 minutes, so the median is 20. That describes the middle order better than 30 when one exceptional order has pulled up the mean. You still investigate the 120-minute order separately before making a delivery promise.

Median for ten orders = (5th sorted time + 6th sorted time) ÷ 2

Why bring in the mode?

Line C’s supervisor asks which packing time showed up most often. Its mean and median both say 30 minutes. Count the times and you see two modes, 25 and 35 minutes. The packing log shows that five orders used standard wrapping and five used gift wrapping. That gives the supervisor a reason to check whether the two order types need separate planning times.

Mode = the value or values that occur most often

LineMeanMedianModeWhat the data tells you
A303030One time describes the whole batch.
B302020Most orders were quick. One exception consumed a large share of the total time.
C303025 and 35A single planning time hides two common routines.

These are all measures of central tendency. The owner looks at the table and asks the next question: “If we plan for 30 minutes, how far off are we likely to be in a batch like this?”

2. How far off is the plan?

On Line C, five orders finish five minutes early and five finish five minutes late. Add those signed differences and you get zero, even though every order misses the plan by five minutes. Count the size of each gap in either direction and the mean deviation from 30 is 5 minutes. Line A has the same mean, but its mean deviation is zero.

  1. Choose a reference time
  2. Measure each absolute gap
  3. Average the gaps

Mean deviation from c = (|x₁ − c| + ⋯ + |xₙ − c|) ÷ n

Here x is an order’s time, n is the number of orders and c is the reference. It could be the mean, median, mode or a time you use for planning.

Which reference is closer to Line B’s orders overall: 30 or 20?
ReferenceGapsMean deviationWhat the data tells you
30 minutes · meanNine gaps of 10 and one of 90(90 + 90) ÷ 10 = 18 minA 30-minute estimate is already ten minutes off for nine parcels.
20 minutes · median and modeNine gaps of zero and one of 100(0 + 100) ÷ 10 = 10 minTwenty fits nine parcels exactly and lowers the average error. The exceptional order still needs its own explanation.

You can now quantify how far a packing plan was off on average. The owner brings you Line D’s record. “What if the same average gap comes from a few orders being much later?”

3. The orders that missed the courier

Line D also averages 30 minutes and has a mean deviation of 5 minutes. Before you give it the same assessment as Line C, check which parcels actually finished after the courier arrived at minute 37.

Line C25252525253535353535
Line D20202530303030354040
Packing times in minutes. Line C got all ten parcels ready by minute 37. Two on Line D were still being packed.

Mean deviation gives an ordinary average of the gaps. Standard deviation gives the bigger gaps extra weight by squaring each one. On graph paper, a ten-minute gap forms a square with four times the area of a five-minute gap.

Early and late gaps of the same size contribute equally. Bigger gaps contribute disproportionately more.
  1. Subtract the mean
  2. Square each deviation
  3. Average the squares Variance
  4. Take the square root Standard deviation

Variance = Σ(xᵢ − x̄)² ÷ n
Standard deviation = √variance

The mean gives the smallest total squared gap among possible reference times. Here you divide by ten because you’re describing the full batch. The square root puts the result back in minutes.

MeasureLine CLine DWhat the data tells you
Mean30 min30 minThe average packing time is the same.
Mean deviation5 min5 minLine D’s smaller and larger gaps average out to the same value as C’s uniform gaps.
Variance250 ÷ 10 = 25 min²450 ÷ 10 = 45 min²The four ten-minute gaps on D have a larger effect once squared.
Standard deviation5 min√45 ≈ 6.7 minD has more spread once bigger gaps get extra weight. Investigate that line before assuming both can meet the same pickup.

Use standard deviation to flag wider variation. Count orders past the cutoff to measure missed pickups. Early and late orders widen the standard deviation alike, while only late orders miss the courier.

4. Is six minutes a lot for this service?

The shop now wants to compare standard orders on Line C with a gift-box service on Line D. Gift boxes have an extra 40-minute personalisation step for every parcel. Add 40 minutes to each of Line D’s times. Its mean rises to 70 minutes and its standard deviation stays about 6.7 minutes.

“The gift boxes vary by more minutes,” the owner says. “Does that mean they’re less consistent for a service that normally takes much longer?”

Coefficient of variation (CV) = (standard deviation ÷ mean) × 100%

Compare variation in minutes and as a share of the mean
ServiceMean timeStandard deviationCVWhat the data tells you
Standard · Line C30 min5 min16.7%Smaller spread in minutes, larger spread relative to the usual service time.
Gift boxes · Line D plus personalisation70 minAbout 6.7 min9.6%Larger spread in minutes, smaller spread relative to its longer service time.

The percentages use unrounded values. A lower CV means less variation relative to the mean. The gift boxes take longer and vary more in minutes. If you’re reserving time before a pickup, look at the times in minutes and count orders past the deadline. If you’re comparing consistency across services whose average times differ, CV puts each spread in context.

In your report to the owner, you keep the average for labour planning, the median for a typical order and the modes for common routines. You add mean deviation for how far the plan was off, standard deviation to flag larger gaps, and CV to compare services on different time scales. Beside the courier schedule you write the number it needs most: how many parcels actually missed pickup.