Budget by Payback Horizon, Not by Channel: A Startup Growth Budget Framework

Most startup marketing budgets are built as a list of channels with a number next to each one. Paid search, content, events, tools, contractors. The list is easy to build and easy to approve, and it hides the single decision that actually determines whether the plan works: how long you are willing to wait for each line to pay you back.

A channel list makes a twelve-month asset and a two-week campaign look like the same kind of purchase. They get reviewed on the same monthly dashboard, compared in the same table, and cut in the same meeting. That is how a content program dies at month four, three months before it would have started working.

This post is a different way to lay out the same money. It is a framework, not a benchmark, and it is labeled that way throughout. If you want the “how much in total” question answered with cited survey data first, start with our Series A marketing budget benchmarks and come back here for the split.

The three horizons

Every marketing line item you can name belongs in one of three buckets, defined by when the money comes back.

That last row is the whole argument. The three horizons fail on completely different schedules, which means a single monthly review will always flatter horizon 1 and always starve horizon 3.

Why channel-based budgets mislead you

Two reasons, and the second is the expensive one.

Channels span horizons. Paid search is horizon 1 on [competitor] alternative and horizon 3 when you are buying category terms nobody is searching with intent yet. A webinar is horizon 2 for your list and horizon 3 for the audience your partner brings. LinkedIn is whichever one you are using it for this month. Budget the job, not the logo on the invoice.

Channel lists invite the wrong comparison. Put content and paid search in the same table with a cost-per-lead column and content loses every quarter for a year, then wins for the following four. If your budget format forces that comparison, your budget format is making the decision for you. This is the same failure we describe in demand generation vs demand capture, where running both motions through one last-touch model quietly defunds the thing producing your branded search.

The horizon-3 evidence most plans ignore

The strongest argument for separating the horizons is what the data says about how long the long one actually takes.

Ahrefs studied 1 million random URLs and 1.3 million keywords, in research updated in May 2025. The findings that should reshape your budget:

• 1.74% of newly published pages rank in Google’s top 10 within a year, down from 5.7% in their 2017 study.
• 72.9% of pages in the top 10 are more than three years old, up from 59% in 2017.
• The average number one ranking page is five years old, up from two years in 2017.
• Only 13.7% of top 10 results were under a year old.

Read that against a typical startup content plan: twelve posts, six months, reviewed monthly on organic sessions. The plan is being measured on a clock roughly one fifth the length of the process it is measuring. It will look like a failure at every single review, right up until it is cut.

How to size each horizon

There is no published people-versus-programs-versus-horizon split for Series A companies. We looked; the only published allocation splits come from Gartner, whose sample skews to very large enterprises, and from agencies selling the services being allocated. So this section is our framework, clearly labeled, built on three rules rather than three percentages.

Rule 1: fund horizon 1 to its ceiling first. Horizon 1 is finite and you can find its ceiling empirically. Keep buying high-intent terms and fixing the conversion path until the marginal cost per qualified conversation exceeds what you will pay. There is no equivalent stopping rule anywhere else in the budget, which is exactly why the other two horizons are so easy to overspend on.

Rule 2: give horizon 3 a standing allocation, not a variable one. The size matters far less than the stability. A modest content and positioning investment that runs uninterrupted for two years beats a generous one that gets cut in the first soft quarter and restarted eighteen months later, because the restart does not resume from where you stopped. If you cannot protect the line, shrink it until you can. What belongs in this bucket, and what to rent instead, is the subject of the compounding assets post. If someone is arguing for a brand budget here on the basis of the 60/40 rule, read this first.

Rule 3: let horizon 2 absorb the remainder, and re-underwrite it quarterly. Horizon 2 is where most of the flexible money should sit, because it is the only horizon where a quarterly review can actually tell you something true.

A worked shape for a first marketing leader with one number to hit, presented as an illustration rather than a benchmark:

The discipline in that table is entirely in the horizon 3 column, and it is the hardest column to hold. It will be the least defensible line in every quarterly review for the first year.

Give each horizon its own instrument

Three horizons on one dashboard is how the wrong one gets cut. Give each its own measurement, and never rank them against each other in the same table.

Horizon 1: traced attribution. The path is short and observable. Cost per qualified conversation, win rate by source, and time to first response are legitimate numbers here and you should hold them tightly. Speed to lead is the single highest-leverage number in this bucket, because the demand is already paid for by the time the buyer arrives.

Horizon 2: cohort tracking. Not last touch. Follow the group of people who entered a sequence or attended an event and watch what that cohort does over two quarters against a comparable cohort that did not. It is coarse, but it is the right shape of measurement for a two-quarter payback.

Horizon 3: leading indicators plus self-reported source. Branded search volume, direct traffic, and an open text “how did you hear about us?” field on your demo form. Read the answers qualitatively, in the words people use. This is the only place a podcast mention or a conference talk ever shows up. Our post on traced vs modeled attribution covers why forcing this bucket into an attribution model produces a confidently wrong number.

The four failure modes this framework prevents

1. Killing horizon 3 at month four. The most common and most expensive. The Ahrefs numbers above are the defense: bring them to the meeting before the meeting happens.
2. Treating horizon 1 as unlimited. Capture demand is finite. Once you have bought all the intent worth buying, more spend in horizon 1 buys progressively worse traffic, and the dashboard keeps looking fine because the average moves slowly.
3. Funding horizon 2 with everything. Middle-horizon programs feel productive, produce visible artifacts, and are the easiest to justify. They are also where budgets go to look busy. Cap it deliberately.
4. Rebalancing on the calendar instead of on evidence. Quarter boundaries are an accounting artifact. Rebalance when a number moves, not when a quarter ends.

Two related traps sit just outside the horizon split and cause the same damage. One is reviewing any horizon before its results can exist, which is pipeline lag and is worth calculating before you set a review cadence. The other is spreading a fixed amount of execution capacity across too many channels inside a single horizon, covered in when to add your second growth channel.

The constraint that is usually not money

Worth saying plainly, because it changes what the budget is for. At a company with one or two marketers, the binding constraint on all three horizons is rarely the media budget. It is production capacity: who writes the sequence, builds the landing page, drafts the twelve posts, and answers the buyer who shows up at 9pm with a question. Our marketing team of one playbook covers the prioritization side of that, and martech stack consolidation covers the swivel-chair tax that eats the hours you do have.

This is the shape Marqeable is built around, and we will scope the claim honestly. On the generate side, campaigns draft the emails, social posts, blog pieces, and images for a full plan with you approving every piece, which is what makes a sustained horizon 3 cadence survivable for a team of one. Automations run the horizon 2 follow-up so a sequence does not depend on someone remembering. On the convert side, AI website chat answers in-market visitors from your own business information instead of handing them a form, replies land in a conversations inbox alongside SMS, and attribution ties revenue back to the exact message so each horizon gets measured on its own instrument. We are in private beta with a small early cohort, so weigh that accordingly.

When this framework does not apply

• Pre product-market fit. If you are still finding out who buys, everything is horizon 1 and everything is an experiment. Budget by experiment instead, using the sizing in how many growth experiments you can afford.
• Pure PLG with self-serve checkout. Your payback horizons compress dramatically and the middle bucket mostly disappears into product.
• A company with under two quarters of runway. Horizon 3 is a luxury that assumes you are here in eighteen months. Fund the short horizon and be honest with yourself about why. Growth targets you can defend without torching runway covers the runway math.
• Enterprise motions with multi-year cycles. Your horizon 1 is somebody else’s horizon 2. Shift every band right and keep the structure.

Frequently asked questions

How should a startup allocate its marketing budget?

Allocate by payback horizon rather than by channel. Sort every line into money back in 0-90 days, money back in 2-4 quarters, and money back in 12 months or more. Give each bucket its own success metric and its own review clock. Fund the short horizon to its ceiling first because it is finite and measurable, protect a standing allocation for the long horizon so it survives quarterly noise, and let the middle absorb the remainder. This prevents the most common budgeting failure: judging a twelve-month asset on a ninety-day dashboard.

What is the 70-20-10 marketing budget rule?

It suggests 70% of budget on proven activity, 20% on emerging bets, and 10% on experiments. It is a reasonable mental model with no published dataset behind it for B2B startups, and it splits on confidence rather than on timing. Those are different problems. A proven channel and an unproven one can both pay back this quarter, and both can take a year. Splitting by payback horizon gives you a review schedule as well as an allocation, which is the part that actually protects long-term work.

How long does content marketing take to pay back?

Longer than most budget cycles allow. Ahrefs’ study of 1 million URLs and 1.3 million keywords, updated May 2025, found only 1.74% of newly published pages reach Google’s top 10 within a year, 72.9% of current top 10 pages are more than three years old, and the average number one page is five years old. Fund organic content on a twelve-month-plus clock, or do not fund it and be honest about the choice.

What percentage should go to long-term marketing programs?

No published Series A benchmark exists for this split, so treat any percentage as a framework rather than data. The stability of the long-horizon line matters more than its size. A modest allocation that runs uninterrupted for two years outperforms a large one that gets cut in the first soft quarter, because restarting does not resume from where you stopped.

The bottom line

A channel list tells you where the money goes. A horizon split tells you when to expect it back, which is the only thing that makes a marketing budget reviewable.

Fund the 0-90 day horizon to its ceiling, because it is finite and you can find that ceiling empirically. Give the 12-month horizon a small, boring, protected allocation and then leave it alone through four budget conversations, because the Ahrefs data says the clock you are working against is measured in years and the average top ranking page is five years old. Let the middle horizon carry the flexible money and re-underwrite it quarterly on cohort evidence.

Then measure each one on its own instrument and never put them in the same comparison table. The horizon that loses that comparison is always the one that was going to compound.

See it live: Marqeable’s campaigns draft the long-horizon content a small team could not sustain by hand, automations run the middle-horizon follow-up, AI website chat answers the in-market buyer in the short horizon, and attribution ties revenue back to the exact message.

Marqeable runs your campaigns, answers every visitor, text, and email in seconds, and turns them into booked jobs and meetings - even at 9pm on a Saturday. We’re in private beta with a small early cohort. Get early access