When the board says extend runway, marketing is usually the first budget line people look at, because it is large, variable and easy to switch off. Switching it off is also how companies create a pipeline hole that shows up as missed revenue two quarters later, exactly when they need to show traction for the next raise. I have helped founders make these cuts, and the goal is never simply to spend less. It is to spend less while keeping the pipeline that pays for the next six months. Here is the order I cut in.
Start with a two-week spend audit, not a percentage
The worst way to cut marketing is a flat percentage across every line. It treats your best channel the same as your worst, and it usually protects whatever has the loudest owner. Before cutting anything, I spend a short, focused period mapping every marketing cost to what it produces. List every recurring cost: media by channel and campaign type, agencies and freelancers, software subscriptions, events, content production, sponsorships and headcount. For each, write down the last two quarters of output in revenue terms: pipeline created, opportunities, customers or, for D2C, contribution margin after returns. Then sort the list into three groups. Group one: spend you cannot connect to pipeline or revenue at all. Group two: spend that produces pipeline, but at a cost per opportunity or cost per customer well above your average. Group three: spend that reliably produces pipeline at an acceptable cost. This audit also exposes how much of your budget is fixed versus variable, which matters because tools and retainers often have notice periods. Start those notice clocks first, since savings there take the longest to land. The ad spend waste calculator linked below is a quick way to size the obvious leakage.
Cut first: spend with no line to revenue
Group one goes first, and in most companies it is bigger than expected. Typical items: software nobody logs into, overlapping tools that do the same job, such as two enrichment providers or two schedulers; agency retainers for channels that have not produced attributable pipeline in two quarters; sponsorships and events chosen for visibility with no follow-up process; brand awareness campaigns with no measurement plan; and content production for formats nobody distributes. Also look for duplicated vendors. I regularly find a company paying one agency for paid social, another for search and a freelancer for landing pages, with nobody joining up the results. Consolidating saves money and usually improves performance. The test I apply is simple: if this line disappeared tomorrow, which metric would we expect to fall, and by when? If nobody can name one, it goes. Be careful with one category here. Some brand and content spend genuinely drives pipeline but is poorly measured. If you suspect that, do not keep it on faith; pause it in a measurable way, for example in some regions only, and watch what happens to branded search and direct demo requests. That is a cheap way to find out whether the spend was doing real work.
Cut second: the least efficient increments, not whole channels
Group two is where judgment matters. The instinct is to kill the channel with the worst average CAC. The better move is usually to trim the least efficient part of each channel. Paid channels have diminishing returns: the first portion of budget buys the cheapest, highest-intent customers, and each extra increment costs more. Andreessen Horowitz makes the same point about paid CAC: costs typically rise as you try to reach a larger audience. So cutting the top increment of a good channel often loses fewer customers than the budget reduction suggests. In practice: in search, cut broad match and generic terms before exact-match brand and high-intent product terms. In paid social, cut cold prospecting audiences and broad awareness objectives before retargeting people who already engaged and conversion campaigns that work. In LinkedIn, narrow to your best-fit accounts instead of broad job-title targeting. Watch cost per qualified opportunity, not cost per lead, while you do it, because cheap leads that never convert are the first thing to go. Make the cuts in steps, wait long enough for the platform to settle, and measure what actually happened to pipeline before cutting again.
Protect: demand capture, follow-up and tracking
Three things I protect almost every time. First, demand capture: the spend that reaches people already looking for what you sell. That usually means brand search, high-intent non-brand search and retargeting of engaged prospects. These buyers are close to a decision, and if you disappear, a competitor answers. Second, follow-up on demand you already paid for. Leads in the CRM, open opportunities, past trials, churned customers who might return and stalled deals cost nothing to acquire again. Email nurture, sales follow-up discipline and lead routing are the cheapest pipeline you have, and in a cut they deserve more attention, not less. If lead response times are slow, fixing that is worth more than any media budget. Third, measurement. It is tempting to cut the analytics contractor or the server-side tracking bill, but during cuts you need to know precisely what is still working. If you lose conversion tracking, the platforms optimise worse and you lose the ability to see the impact of your own cuts. Also protect one experiment slot, even a small one, so you keep learning which channel to scale when budget returns.
People, agencies and the order of operations
Headcount is the most painful line and the slowest to rebuild, so I sequence it last among marketing cuts and decide it with the founder, not by formula. Before reducing your own team, look at agencies and contractors. If an agency is doing work your team could absorb with fewer channels to run, ending that retainer often saves as much as a role. Check notice periods and make sure you control the ad accounts, analytics and pixels before giving notice; my agency switching checklist covers the handover. When you do reduce scope, reduce channels at the same time. A smaller team running the same number of channels does all of them badly. Pick the two or three channels in group three and do those well. Finally, model the effect before you act. Take your cut plan, estimate the pipeline each removed line was producing, and run it through your reverse funnel to see the revenue impact two quarters out. Show that to the CEO and CFO next to the runway extension. The growth runway simulator linked below helps frame that trade-off. A cut that adds three months of runway but removes a quarter of next year's pipeline may not be the cut you want.
Sources
Andreessen Horowitz, 16 Startup Metrics (paid CAC typically rises as you reach a larger audience): https://a16z.com/16-startup-metrics/
FAQ
Rarely. Pausing everything usually removes the cheapest, highest-intent demand along with the waste. Cut spend with no link to revenue first, then the least efficient increments of each channel, and protect brand and high-intent search.
Unmeasured brand spend often goes early, but test before cutting entirely. Pause it in some regions and watch branded search and direct demo requests. If they hold, the cut was safe.
Invest attention in demand you already paid for: CRM follow-up, nurture, fast lead response and stalled deals. Keep tracking intact so you can see the effect of each cut, and make cuts in steps.