How Foodhub moved marketing ROI from −18% to +22% in twelve months

The situation

Foodhub is a food-ordering marketplace operating across the UK, United States, Australia and two further markets. By the time I joined as VP Marketing — the first senior marketing hire, reporting directly to the Chief Revenue Officer — the business had grown fast enough that marketing spend had scaled but the structure underneath it had not.

The headline number was blunt: marketing was returning −18% on spend. Every pound going out was coming back as less than a pound. That had continued for long enough to become normal, and the internal explanation had settled into a familiar standoff. Marketing believed the leads were fine and sales was not converting them. Sales believed the leads were unqualified. Neither position could be tested, because nobody had written down what a qualified customer was.

The diagnosis

I spent the first three weeks doing almost nothing visible: reading the pipeline data, sitting with the sales team, talking to restaurant partners, and mapping where money entered and left the system. Three structural problems came out of it.

One funnel serving two completely different businesses

Foodhub is a two-sided marketplace. Restaurants are the supply side — a B2B sale with a long consideration cycle, a relationship-driven renewal, and meaningful revenue per account. Consumers are the demand side — a fast, low-consideration, high-frequency transaction. These are not variations of one customer. They are two businesses.

Both were being served by a single acquisition budget and a single funnel definition. The predictable result: spend flowed toward whichever side was cheaper to acquire in a given week, rather than whichever side was actually constraining growth. Consumer acquisition looked efficient in isolation while the restaurant base quietly eroded underneath it.

Five markets running on five different assumptions

Each market had built its own working definitions — of a lead, of an active customer, of a churned account. Roll-up reporting therefore added together numbers that did not mean the same thing. The board was reading a group figure assembled from five incompatible local ones.

No shared definition of a qualified opportunity

This was the root cause and the reason the marketing-versus-sales argument was unwinnable. Marketing was measured on volume of leads. Sales was measured on closed revenue. Nobody owned the conversion between them, so nobody was accountable for the only number that mattered.

What I did

1. Split the commercial function into two funnels

Restaurant acquisition and consumer acquisition were separated entirely: different budgets, different targets, different teams, different metrics. Each side got its own definition of a qualified customer, written down and agreed with sales before any spend moved.

2. Restructured the team around markets, not channels

The team had been organised by channel — paid, CRM, content, brand — which meant nobody owned the outcome in any single market. I reorganised around market ownership with shared central capability, so each market had one person answerable for its number and a common toolkit to deliver it.

3. Attacked churn on both sides, differently

Retention was the fastest available margin. The two sides needed opposite interventions:

These are two separate numbers and they should never be averaged. Averaging them would have concealed both problems and sent the budget to the wrong place.

4. Connected every channel to a revenue target

No channel kept its budget on the strength of an activity metric. Every line of spend was tied to a pipeline or revenue target that sales had agreed to. Anything that could not be attributed was stopped rather than optimised. That single decision produced most of the early ROI movement — the gain came from what we stopped, not from what we added.

The results

Nick is an asset to any business, a strong commercially focused marketer who is driven to deliver success. He is a pleasure to work with and a popular member of the team.

Simon FarmerChief Revenue Officer, Foodhub

What transfers to other businesses

Foodhub was a marketplace, but nothing about the diagnosis was marketplace-specific. Three things transfer directly:

• Negative ROI is rarely a creative or channel problem. It is usually a definition problem. If two functions cannot agree what a qualified customer is, no amount of optimisation will fix the number between them.
• Segments that behave differently must be measured differently. One blended churn figure across two customer types hides both problems and misdirects budget toward the easier one.
• The fastest ROI gains come from stopping, not starting. Cutting unattributable spend moves the ratio immediately and costs nothing. Most turnarounds begin here.