Case 059Industry structure and moatsCore
Mandivar Marketplace, a B2B marketplace, raises its take rate from 2% to 3.5% and supplier churn rises from 5% to 8%. Does the network hold, and what is the revenue effect?
1The situation
Mandivar Marketplace connects businesses buying industrial supplies with 50,000 suppliers. Each supplier sells about Rs 40 lakh a year through the platform, so gross merchandise value, GMV, is Rs 20,000 crore and at a 2% take rate revenue is Rs 400 crore.
Mandivar raises the take rate to 3.5%. Annual supplier churn, the share of suppliers who leave each year, rises from 5% to 8%. About 2,500 new suppliers join each year, which at 5% churn held the base steady at 50,000. Supplier gross margins in this category are around 15%.
2Your task
Work out the revenue effect in year 1 and in the long run, and decide whether the network is holding.
Quick check
If sign-ups stay at 2,500 a year, where does the supplier base settle at 8% churn?
Worked solution
Try it on paper, then open one step at a time.
30-second answerThe answer to give first
Revenue rises about 70% in year 1, from Rs 400 crore to Rs 679 crore, but the long-run gain is only about 9%. At 8% churn and 2,500 sign-ups a year, the supplier base slides towards 31,250. Churn rose only 3 points for a 75% price rise, so pricing power is real, but the network is getting thinner. It holds only if sign-ups do not fall.
Step 1What does a price rise reveal about a network?
A popular weekly market raises stall rent by 75%. If only a handful of stallholders leave, the market clearly has something they cannot get elsewhere: the crowd. A network's pricing power shows in how little churn a price rise causes. Here a 75% rise in the take rateThe share of each transaction’s value that the platform keeps as its fee. moved annual churn from 5% to 8%, a mild reaction, which says suppliers value Mandivar's buyers.
Step 2Why is year 1 the wrong year to judge the revenue effect?
In year 1 only 3% more suppliers leave, so GMV falls to Rs 19,400 crore and revenue is Rs 679 crore. But churn repeats every year. A supplier base settles where the suppliers who join equal the suppliers who leave: 2,500 divided by 8% is 31,250, down 37.5%. At that size GMV is Rs 12,500 crore and revenue Rs 438 crore, only 9% above the old Rs 400 crore.
| N* | the supplier count where joiners equal leavers |
| sign-ups | new suppliers joining each year, 2,500 |
| churn | share of suppliers leaving each year |
Step 3So does the network hold?
On the numbers, narrowly. Churn at 8% is under the 8.75% breakeven, and a 3.5% fee takes about 23% of a supplier's 15% gross margin, which is still cheaper than most offline selling routes. But the test is the next data point, not this one: watch whether sign-ups hold at 2,500 and whether the suppliers leaving are small ones or the large ones buyers come for. Losing the biggest suppliers shrinks GMV faster than the count suggests and weakens the buyer side too, which is how a network unravels.
The limitation of the model is that every supplier is assumed to be the same size and sign-ups are assumed not to react to the fee. Both are the first things to test with Mandivar's cohort data.
Where candidates lose it
The usual loss is multiplying year 1 GMV by the new take rate, declaring 70% revenue growth, and stopping. Churn is a rate that repeats every year; the base keeps shrinking until joiners equal leavers.
The second is the reverse: seeing churn rise and declaring the network broken. A 75% price rise that costs 3 points of churn is evidence of pricing power, not weakness.
What the interviewer asks next
- How would you test whether the suppliers leaving are the large ones?
- What would buyer-side data tell you that supplier churn cannot?
- At what take rate does long-run revenue peak, if churn rises 2 points for every 1 point of take rate?
Company names and figures are illustrative.