Hyrax · GTM · September 2, 2026

Pre-Revenue Budget Model

Replacing the ICONIQ revenue chain with a method that works at zero revenue. Built entirely on published benchmarks and market capacity.

Recommendation: ask for $400,000 of 12-month program budget, with a further $120,000 released only against a measured cost per paying account. Published method returns $378K–$486K for the amount of net-new ARR this budget can realistically produce, so $400K survives cross-examination in a way the $500K social ask does not.

$400K
Base 12-month program budget, non-headcount
$520K
With the conditional media reserve released
$45.5K
Monthly all-channel paid ceiling the category can absorb at target efficiency
~$1M
Net-new ARR $400K buys at median B2B SaaS efficiency
Every figure on this page comes from a published benchmark, a competitor estimate, or arithmetic on the two. Nothing here is derived from internal spend or internal performance data, so the model can be shared and defended without disclosing either.

1The problem with the current ask

The deck chains five ICONIQ percentages off a $10M year-one ARR target: 56% marketing OpEx × 55% programs × 63% online × 26% social × 25% X share = $126K for X, inside a $505K social envelope.

Every percentage in that chain is real and the arithmetic is right. The instrument is wrong for this decision, for one reason: ICONIQ's 56% is marketing OpEx measured against realized revenue for companies already operating at $1–10M ARR (Marketing Budgets & Productivity, p.10, data as of 2024-10-29). Pointed at a target instead, it computes what a $10M-revenue company spends. It does not compute what a pre-revenue company should spend to become one.

The question it invites

The chain says Hyrax's marketing OpEx should be $5.6M. The company is not funding $5.6M. Asking for the social slice of a plan that does not exist gives the slice no more authority than the whole, and the first reader who follows the chain backwards asks where the other $5.1M sits. The ask needs a denominator that exists.

The category cannot absorb the ask

Independent of any spending history, the amount of paid media this category can absorb at target efficiency is computable from search volume, published CPMs and published CPAs. It comes to roughly $45.5K/month across every channel, of which the social channels account for about $23K.

ChannelRecommendedCeilingWhat sets the ceiling
Google Search (non-brand + conquest)$10,500$15,000~12,400 four-country non-brand searches/month exist in the category. At 80% impression share and 5.11% B2B SaaS search CTR, that is ~510 clean clicks/month. Conquest terms add ~24,700 searches at low CPCs.
Reddit (6 core developer subs)$4,500$9,00013.1M members in target subs. Inventory could take $45K+, but published B2B CPAs of $50–150 cap the quality slice near $300/day.
X (developer-interest)$2,500$6,000$8–15 B2B tech CPM. Reach is cheap; measurement is the constraint.
LinkedIn (ABM list + thought leader)$2,000$5,000A named-account list of this size saturates at $1,000–3,600/month before frequency fatigue.
YouTube / Demand Gen$0$7,500Step function. Google requires ~50 conversions to exit learning; at the $130 devtools CPL that is $6,500/month minimum or nothing.
Retargeting (cross-channel)$1,500$3,000Retargeting runs efficiently at 10–15% of prospecting spend, and a warm pool caps out at sane frequency.
Total$21,000$45,500

The $500K social ask is $41.7K/month on social alone, roughly 1.8× the social ceiling and 92% of the ceiling for every channel combined. Any reader who checks the category math reaches that on their own, so the model should reach it first.

2The method that replaces percent-of-revenue

Published practice offers three denominators that exist before revenue does. One of them can be run entirely from public data.

MethodPublished rangeStatus
A. Percent of capital raised 10–20% of the round to marketing; $50K–$250K/yr typical at seed Optional cross-check. Requires a finance input.
B. Percent of burn 15–30% of burn Optional cross-check. Requires a finance input.
C. CAC-backed from net-new ARR Marketing = net-new ARR ÷ CAC ratio × 35–45%. Median new-customer CAC ratio is $2.00 of sales and marketing per $1 of net-new ARR. Programs are 54% of marketing below $25M ARR. Runs on public data alone. This is the primary method.
Method C, worked

$1M of net-new ARR × $2.00 CAC ratio = $2M of sales and marketing. Marketing takes 35–45% of that: $700K–$900K all-in. Programs are 54% of marketing below $25M ARR: $378K–$486K. The $400K base sits inside that band.

Run the same math on the $10M target and it returns $3.8M–$4.9M of program spend, which corroborates ICONIQ's $5.6M rather than contradicting it. Two independent methods agree that $10M of ARR costs several million in marketing. Reconciling the target with the budget is the decision in front of the company; the budget number is downstream of it.

3Working backwards from $1M/month

At $1M/month ($12M ARR) and the published $30/user/month floor price, the required account count is decided by seats per account. Cost per free signup and free-to-paid conversion are taken from published category benchmarks.

$100
Planning cost per free signup. Published range $42 best case to $150 typical for developer tools
3% / 7%
Developer-tools freemium conversion. Category median 2–4%, top performers 7%+
$30
Per user per month, the published pooled floor price
1 of 8
Scenarios below that sit inside the $45.5K/month category ceiling
Seats per paying account ARR per account Accounts for $12M Signups/mo at 3% Media/mo Signups/mo at 7% Media/mo
10 seats$3,6003,3339,258$926K3,968$397K
20 seats$7,2001,6674,631$463K1,985$199K
50 seats$18,0006671,853$185K794$79K
100 seats$36,000333925$93K396$40K
Media cost assumes $100 per signup holds at volume, which it does not above the category ceiling. Every cell above $45.5K/month is therefore optimistic, not conservative.
The finding

Seven of the eight scenarios require more monthly media than the category can absorb at target efficiency, several by an order of magnitude. Only one combination works: 100-seat accounts converting at top-quartile rates, and even that needs $40K/month against a $45.5K ceiling.

So the path to $10M runs through seats per paying account and free-to-paid conversion, not through media budget. No budget reaches the target at small account sizes, because the demand pool does not exist to buy. That reframes the submission from "how much do we spend" to "what does each dollar buy," which is the version of the conversation the company can win.

4Recommended 12-month budget

LineMonthlyAnnualBasis
Paid media — search core (non-brand + conquest)$10,500$126,000Category capacity math in section 1
Paid media — Reddit, X, LinkedIn ABM$9,000$108,000Half of each channel's published quality ceiling
Paid media — retargeting$1,500$18,00010–15% of prospecting spend
Contractors (technical writers, design, video edit)$5,000$60,000Seed norm $50–80K/yr. Agencies and outsourcing are 15–25% of marketing spend (ICONIQ pp.17–18)
Video production$2,500$30,000Seed benchmark $12–25K, set above it to fund a monthly recording cadence plus one product demo
Martech and tooling$1,500$18,000Practical seed stack $12–18K/yr; 8–12% of budget when the CRM also carries founder-led sales
Events and community$1,250$15,000DevOpsDays startup sponsorships $750–999 plus the 3×-the-space-fee all-in rule
PR$833$10,000Project-based launch work only. 2–5% is the pre-PMF share
Experiments and contingency$1,250$15,000Insight Partners: ~10% of media reserved for testing
Base total$33,333$400,000Inside the $378–486K band from Method C
Conditional media reserve$10,000$120,000YouTube re-entry ($6,500 step function) plus scale headroom. Released against a measured cost per paying account
Total if the reserve is released$43,333$520,000Still inside the $45.5K/month category ceiling
Why staging beats a single number

Splitting the ask was the right instinct; the trigger was the missing piece. Tie the second tranche to a measured cost per paying account rather than a revenue target. Cost per paying account is a number marketing controls and can produce inside one quarter, while a revenue target can be missed for reasons that have nothing to do with media performance, killing the tranche unfairly.

5Answers to the specific questions

How much on paid media?

$21K/month recommended, $45.5K/month ceiling. The ceiling is a quality ceiling rather than an inventory one: Reddit's core developer subs alone could absorb $45–107K/month before frequency exhaustion. What caps spend is that published category CPAs sit above any efficient target. Reddit B2B CPA runs $50–150, LinkedIn CPL $80–160, and devtools search CPL $130–143. Each channel has a modest slice that converts efficiently and a long tail that does not.

What channels should we run?

Four. Google Search as the core, Reddit and X as mid-funnel volume, LinkedIn against a named-account list, with retargeting as a layer rather than a fifth channel. YouTube is a separate step-function decision at $6,500/month or zero, because below that the campaign never exits learning and the spend is wasted.

Insight Partners surveyed 100+ portfolio companies in 2024 and found 70% of marketing-sourced pipeline came from four channels — events, SEO, social and paid search — regardless of growth stage or deal size. Their recommendation is to cut channel count and stop programs without measurable pipeline.

We are running about ten channels. What do we cut?

Set the criterion before the cut: cost per paying account, by channel, over one buying cycle. That criterion requires a source tag on every signup and a clean handoff from site to app, so instrumentation comes first and the cut follows within a quarter. Cutting ahead of measurement risks killing the channel that works, and the budget submission should name instrumentation as a gate rather than assume it.

PR: should there be a budget?

Yes, $10K, project-based, held for a launch or a funding announcement. Not a retainer. Boutique technology PR runs $3–6K/month, specialist agencies $7.5–15K/month, mid-size firms $15–25K/month, and a focused launch project $8–15K. Published guidance puts PR at 2–5% of a pre-PMF marketing budget and 5–10% once there is a real launch with customer evidence. PR amplifies a story rather than creating one, and the story is still moving under the platform repositioning.

Video production?

$30K for the year, above the $12–25K seed benchmark because the plan includes a recurring recording cadence rather than one-off assets. Reference costs: a polished SaaS product demo runs $3–15K, short-form editing from supplied footage $75–400 per clip or $1,500–2,500/month for 8–12 clips, YouTube edit-only $100–600 per episode, and a 15–30 second digital brand spot $5–25K. A traditional broadcast-style commercial shoot at $15–75K is the line item to keep declining.

Tools and apps?

$18K for the year. A practical seed stack runs $12–18K and a lean founder-operated one $4–8K; a premature enterprise stack runs $50–100K+ and is the most common early-stage overspend. Gartner's 2025 survey found only 49% of purchased martech capability actively used, and Zylo's 2026 index found 36% of software licenses unused. Send the current list and the overlap pass takes about an hour.

The specific trap here is paid attribution. HockeyStack starts around $1,399/month and Dreamdata around $599/month, and neither one repairs a broken tracking handoff. Fix the tracking first and buy the platform only if volume justifies it afterward.

Contractors?

$60K for the year, the largest flexible line after media. Published US rates: technical and developer content writers $0.40–1.20/word or roughly $90–130/hour, senior B2B SaaS writers $1,500–3,000 for a 2,000-word feature, freelance designers $50–100/hour at mid-level, video editors $30–80/hour, and DevRel consultants $4–8K/month against $180–250K fully loaded for the equivalent employee. Contractors should out-earn new full-time marketing headcount until a channel shows repeatable pipeline.

Decline percent-of-spend agency deals at this budget size. The market rate is 10–20% of ad spend, which prices a $21K/month account at $2–4K/month for work a flat $500–2,000/month freelancer covers.

6Open inputs

The model above is complete and defensible as it stands. These four inputs would tighten it, and each is optional rather than blocking.

  1. Capital allocated to Hyrax, and monthly burn. Adds methods A and B as independent cross-checks on the $400K.
  2. Average seats per paying account. Selects the row in section 3 and settles whether $10M is reachable at all.
  3. The tool list with monthly cost. Turns the $18K line from a benchmark into a real number and produces a cut list.
  4. Contractor and agency commitments. Same, for the $60K line.