Channels and the mix · Brand and performance

Split $1M between brand and performance

Choose a brand-to-performance split for a $1M budget against a 12-month payback board, and set the six-month result that would change it.

casehardExercise · 3 min · 25 of 60

On this page

The CEO and the CFO have brought next year's $1M marketing budget to the board with two different splits, and the board has asked marketing to choose one and defend it.

The situation

The invoicing app's budget for next year is $1M. Use the model from brand versus performance, which counts each customer's lifetime margin, $800, because the splits are compared over three years; the 12-month payback is a separate test. Performance returns lifetime margin in tiers: $2.40 per dollar on the first $300,000 (customers at $333 on the course's $2-click chain), $1.20 on the next $300,000 ($667 a customer) and $0.70 on the next $300,000 ($1,143 a customer). Brand returns $0.40 per dollar in the year of the spend, $0.80 the year after and $0.60 in the third, $1.80 in all; that figure comes from category benchmarks and has never been measured for this company. The CEO wants 60:40, brand to performance, "because Binet and Field". The CFO wants 10:90, because the board judges marketing on payback against a 12-month target, and the company raises again in 15 months with 30 months of runway. At least $100,000 must stay on brand.

The question

Choose the split for next year, say what you tell the board about payback, and name the result in the first six months that would change the split.

A strong answer · Rubric · Where answers go wrong

Write your answer first, then open

A strong answer

Start at the margin, not the ratio. The third performance tier buys $800 customers for $1,143, returning $210,000 for $300,000 at every horizon, so it is cut whatever brand is worth. The CFO's 10:90 funds it. The second tier returns $1.20 over a lifetime, but its $667 customer pays back in month 35: it fails the board's test while still making money.

Brand against the second tier is the real choice. With the budget recurring, a dollar a year earns $3.40 inside 36 months on brand against $3.60 in the second tier; brand wins only in a steady year, $1.80 against $1.20. The 36-month answer flips at a three-year brand return of $1.91, and the assumed $1.80, unmeasured, sits 6% below it.

Choose 40:60: $600,000 on the first two performance tiers and $400,000 on brand, spread over the year and spent in half the regions as a geo test against the other half. It brings $1.24M of lifetime margin from first-year customers against $1.33M for 10:90, $90,000 less, and $4.60M against $4.21M over three years. It also gives the raise in month 15 a measured brand return instead of a benchmark.

To the board, report each tier's payback separately: the first $300,000 buys at $333 and pays back in month 11; the second at $667, month 35, profitable but outside target; brand is a test with a read date, not a payback claim.

If branded search and direct sign-ups in the brand regions are not clearly above the held-out regions by month six, return the unspent $200,000 of brand to runway, not to the third tier. If they are well above, plan 60:40 for the year after.

Rubric

Cuts the $1,143 tier on lifetime arithmetic, independent of the brand question
Separates the second tier's lifetime profit from its failure of the 12-month test
Names brand's unmeasured three-year return as the input the split turns on, with the threshold where the answer flips
Chooses a split and runs the brand money as a geo test with a holdout and a read date
Reports payback to the board per tier rather than as one blended figure
States the six-month result that would move the split in each direction

Where answers go wrong

"60:40, because Binet and Field found it optimal." Their 60:40 is the peak of a curve fitted to award entries, and their own business-to-business analysis put the balance nearer 46:54. The answer adopts an unmeasured return and gives up $250,000 of first-year margin against 10:90 without a test that could justify it.

"10:90, because only performance pays back inside 12 months." It keeps $300,000 a year in a tier that returns $0.70 per dollar for ever, letting the board's metric rank a permanent loss above a delayed gain.

"50:50 as a compromise." A split chosen to settle a disagreement answers neither argument. Nothing in it follows from the returns, and it names no test that would end the argument next year.