PathTailor
FRSign inBuild my path
PathTailor
Build my pathSign inLire en français

Conversion, average order value and revenue per visit

First lesson of the E-commerce Analytics course, free to read: about 15 minutes, 3 exercises and 3 quiz questions. This is the catalogue's base lesson; in the student space, it is rewritten for your profile.

Objectives

  • Break a store's revenue down into sessions, conversion and average order value
  • Compute revenue per visit (RPV) by segment and overall
  • Analyze the distribution of order values (median, extreme values, free-shipping threshold)

Conversion rateRevenue per visitOrder value distribution

Three levers and one summary metric

The DA2 course laid out the decomposition of an online store’s revenue:

revenue = sessions x conversion rate x average order value

To compare segments, a fourth metric is very useful: revenue per visit (RPV), that is, revenue divided by the number of sessions. It combines the store’s two own levers:

RPV = revenue / sessions = conversion rate x average order value

By device

traffic["conversion"] = traffic["orders"] / traffic["sessions"]
traffic["aov"] = traffic["revenue"] / traffic["orders"]          # average order value
traffic["rpv"] = traffic["revenue"] / traffic["sessions"]
Device Sessions Conversion Average order value RPV
mobile 6,000 1.5% €60 €0.90
desktop 3,000 3.2% €90 €2.88
tablet 1,000 2.4% €70 €1.68

Mobile brings 60% of visits but only 34% of revenue: each mobile visit earns three times less than a desktop visit. Gaining half a percentage point of conversion on mobile would be worth 6,000 x 0.005 x 60 = 1,800 euros over the period.

Overall RPV is computed on the totals: 15,720 / 10,000 = 1.57 euros. The average of the three RPVs (1.82) would be wrong, because the segments do not have the same number of sessions.

The distribution of order values

The average order value summarizes a distribution that is often highly skewed: many small orders, a few very large ones.

print(amounts.mean())     # 69.44
print(amounts.median())   # 47.50
print(amounts.max())      # 420 : an exceptional order

A single 420-euro order (a business customer? a data entry error?) pushes the mean more than 20 euros above the median. Before commenting on a rise in average order value, look at the median and the largest orders. If they belong to a different business (professional purchases), handle them in a separate segment rather than deleting them.

The free-shipping threshold

Many stores offer free delivery above a certain amount. Customers adjust their baskets: you often see a buildup of orders just above the threshold, and sometimes orders stuck just below it.

near = amounts.between(free_shipping_threshold - 10, free_shipping_threshold, inclusive="left")
print(near.mean())        # share of orders between 40 and 50 euros

inclusive="left" includes 40 and excludes 50. Here, 44% of orders fall within 10 euros below the threshold: a complementary product suggestion under 10 euros could push them over. But beware: lowering the threshold or adding a suggestion also changes margin and the behavior of other customers. The effect is measured with an A/B test (courses MK1 and ST3), not with a simple projection.

Reading the metrics together

The three levers interact. An aggressive promotion can increase conversion and lower the average order value; a traffic increase from a poorly targeted campaign often lowers conversion. That is why RPV and total revenue are read together with the levers, and always on the same definition of sessions and orders (validated orders, excluding cancellations, with or without tax specified).

Key takeaways

  • RPV = revenue / sessions = conversion x average order value; it compares segments.
  • Overall ratios on sums, never as an average of ratios.
  • Look at the median and extreme values before commenting on the average order value.

Exercises

Solutions and tests stay on the server. In the student space, your code is run and checked, and the tutor gives hints before the solution.

  1. CodeDifficulty 1Python in the browser2 hints

    Conversion, order value and RPV

    Add the columns conversion, aov and rpv to traffic. Compute global_rpv on the totals, best_device (the device with the highest RPV) and mobile_gain: the additional revenue if mobile conversion rose by 0.5 point at constant average order value.

    import pandas as pd

traffic = pd.DataFrame({
    "device":   ["mobile", "desktop", "tablet"],
    "sessions": [6000, 3000, 1000],
    "orders":   [90, 96, 24],
    "revenue":  [5400.0, 8640.0, 1680.0],
}).set_index("device")

# global_rpv = ...
# best_device = ...
# mobile_gain = ...

  2. CodeDifficulty 2Python in the browser2 hints

    Order value distribution

    From amounts, compute mean_value, median_value, mean_without_max (mean without the largest order) and near_threshold_share: the share of orders in [threshold - 10; threshold[.

    import pandas as pd

amounts = pd.Series([22, 35, 41, 44, 47, 48, 49, 52, 55, 60, 75, 38, 46, 49, 30, 420], name="order_value")
free_shipping_threshold = 50

# mean_value = ...
# median_value = ...
# mean_without_max = ...
# near_threshold_share = ...

  3. Multiple choiceDifficulty 21 hint

    Reading RPV

    Channel A converts at 4% with a 30-euro order value; channel B converts at 2% with a 75-euro order value. Which channel earns more per visit?

    • A, because it converts twice as well
    • B: 1.50 euros per visit versus 1.20 for A
    • They are equal
    • Impossible to say without the number of sessions

End-of-lesson quiz

  1. 10,000 sessions, 250 orders, 15,000 euros of revenue. RPV?

    • 60 euros
    • 1.50 euros
    • 2.5%

  2. Why present the median order value next to the mean?

    • Because the mean is sensitive to exceptional large orders
    • Because the median is always higher
    • Out of habit

  3. Many orders between 40 and 49 euros for a free-shipping threshold of 50 euros. What should you do?

    • Lower the threshold immediately
    • Form a hypothesis (complementary product suggestion) and test it
    • Remove free shipping

The rest of the course

  1. Retention and repurchase: delays between orders and maturity15 min
  2. Cohort analysis: retention and revenue by arrival month20 min
  3. Recommendations: products bought together18 min
  4. Store dashboard and RFM segmentation15 min
Build my pathSee the full syllabus

E-commerce Analytics · Intermediate · 1 h 23 min