Dev agency cash flow management 2026

Dev agency cash flow management in 2026 comes down to five rules: take 50% deposits up front, enforce net-15 or net-30 terms (never net-60), keep 3 to 6 months of operating runway, watch your accounts-receivable aging weekly, and treat invoice factoring as the last resort, not a strategy. The agencies that survive a slow Q1 are the ones who priced cash flow into the contract, not the ones who chased it after the fact.

Most dev shops do not fail because the work dried up. They fail because $180,000 of perfectly good invoices were sitting at net-60 with a Fortune 500 client while payroll was due on Friday. Cash flow is the actual product you sell, and most agency owners only learn this after they have to make a list of which engineer to let go.

The agency cash-flow problem in plain numbers

A 6-person dev shop typically runs $80,000 to $140,000 in monthly burn. Salaries (say $9,000 to $15,000 fully loaded per engineer in the US, $4,000 to $7,000 LATAM, $3,000 to $5,000 EU East), plus rent, plus subscriptions, plus the founder draw. The revenue side is lumpy: project work invoiced in milestones, retainers paid monthly, the occasional big rebuild billed in tranches.

The mismatch is structural. Salaries clear every two weeks. Invoices clear in 30, 45, or 60 days. The gap between those two clocks is what kills shops.

Here is what a normal month looks like with mixed client terms.

Day	Inflow	Outflow	Running cash
1	$0	$0	$100,000
5	$0	$42,000 (payroll)	$58,000
15	$25,000 (retainer A)	$0	$83,000
19	$0	$42,000 (payroll)	$41,000
22	$0	$9,000 (rent + SaaS)	$32,000
28	$60,000 (net-30 settled)	$0	$92,000

If the net-30 slips to net-45, you hit Day 28 with $32,000 in the bank and another payroll cycle six days away. That is the moment owners panic-call their bank or their factor.

The 50% deposit rule (and why it is non-negotiable)

A 50% deposit on signing is the single most powerful cash-flow tool an agency has. It does three things at once: it pre-funds the first month of work, it filters out tire-kickers (anyone who refuses is signaling they will fight every invoice), and it sets the precedent that you are a vendor not a creditor.

The structure we recommend for a $60,000 fixed-scope project:

• 50% on signing: $30,000
• 25% at mid-milestone: $15,000
• 25% on delivery: $15,000

For retainers, charge the first month up front and require auto-debit (Stripe, GoCardless, or ACH) on the 1st of every month. Retainers paid in arrears are not retainers, they are receivables wearing a costume.

The only acceptable exception is enterprise or government work where procurement literally cannot cut a deposit check. In that case the deposit gets replaced by a much shorter payment term and a higher rate. Compensate for cash risk with price.

Net-15, net-30, and the dangerous net-60 enterprise trap

Default payment terms in 2026 across US and EU agencies look roughly like this:

Term	Typical client	Cash-flow impact	When to accept
Net-7 / net-15	Founders, seed startups, friends-of-friends	Excellent. Cash lands inside one payroll cycle.	Always; set as your default.
Net-30	Series A/B startups, mid-market SaaS, most agencies	Workable if priced in. Build a 30-day buffer.	Default for most B2B SMB work.
Net-45	Mid-market with slow AP, some non-profits	Tight. Forces a runway buffer of 1.5 months on that client.	Only with a deposit and a rate premium.
Net-60	Enterprise, Fortune 1000, government	Dangerous. You become a free lender for two months.	Only at a 20 to 30% rate premium, and never more than 30% of revenue concentration.
Net-90	Mostly old-school enterprise procurement	Avoid. If accepted, treat as financing not revenue.	Almost never.

The net-60 enterprise trap is real and predictable. A logo client signs at $40,000/month with net-60 terms. You are excited. You staff up two engineers. Month 1 invoice goes out on day 30. Settlement lands on day 90. You have now floated 90 days of payroll for those two engineers ($54,000 to $90,000 depending on cost) before a single dollar arrives. If anything goes wrong, a delay in approval, a disputed line item, a holiday in their AP department, you are underwater.

The rule we tell agency owners: no single client should represent more than 30% of monthly revenue if they pay net-60 or longer. And every net-60 contract needs a late-payment clause (1.5% per month after due date) that you actually enforce.

For more on how to write these clauses without scaring clients off, our breakdown of dev agency contract templates and the gotchas to watch covers the exact language we recommend in the MSA and SOW.

The accounts-receivable aging tracker

Every Monday morning, before the standup, before email, before coffee, the owner or ops lead pulls the AR aging report. This is non-negotiable. The report has four buckets:

• Current (not yet due): informational
• 1 to 30 days overdue: send a polite reminder
• 31 to 60 days overdue: phone call to the AP contact, copy the project sponsor
• 60+ days overdue: escalate to founder/CEO, freeze new work, send the formal demand letter

Most accounting tools (QuickBooks, Xero, FreshBooks, Wave) have this report built in. If you are running on spreadsheets it takes 20 minutes to build. The point is not the tool; the point is the weekly cadence. Invoices that get a phone call inside the first 30 days of overdue get paid 3 to 5 times faster than invoices that get a polite email at day 45.

A working AR target for a healthy agency: Days Sales Outstanding (DSO) under 35. If your DSO creeps above 45, you have a collections problem, not a sales problem. Hire a part-time bookkeeper or AR clerk (around $1,000 to $2,000/month) before you hire your next engineer.

Runway: 3 months is survival, 6 months is freedom

Operating runway is months of cash on hand divided by monthly burn, ignoring future revenue. The targets are not negotiable:

• Under 1 month: emergency. Stop hiring. Cut everything optional. Have the conversation with senior staff about deferred comp.
• 1 to 3 months: survival mode. Aggressive collections, defer founder draws, no new fixed costs.
• 3 to 6 months: healthy. You can negotiate from strength, walk away from bad deals, weather a 60-day dry spell.
• 6+ months: freedom. You can take strategic risks (a new vertical, a productized offering, hiring a sales lead).

Most agencies operate in the 1 to 3 month band because they reinvest every dollar into headcount. That is fine when sales pipeline is strong. The moment pipeline softens for two consecutive months (and it will, every agency owner has a bad Q1), the 1-month buffer becomes a 1-week emergency. Build to 3 months, then push to 6 before you spend the next $50,000.

Bench cost discipline

Bench is the most expensive line item in a dev agency: a paid engineer with no billable client. A senior engineer on the bench at $11,000/month fully loaded is burning $367 per day, every day, that they are not billed. Five days on the bench eats a $1,800 margin. Ten days eats more than most agencies make on a typical SOW.

The discipline rule: every benched engineer triggers an action within 48 hours. Either reassigned to a billable project, put on internal R&D (with a budget cap and an end date), or, if benching repeats and pipeline is soft, contract terminated or hours reduced. Sentimentality on the bench is one of the top three killers of agency cash flow.

This is also where the booked-engineer model wins structurally over full-time hires. When work is spiky, you book capacity for the spike (every engineer on Cadence is AI-native by default, vetted on Cursor, Claude Code, and Copilot fluency before they unlock bookings) and release it when the spike ends, with no severance, no notice period, no bench cost. The 12,800-engineer Cadence pool means a senior is available inside 48 hours when you need them, and gone the same week when you do not.

The agencies who run a hybrid model, a small permanent senior core plus booked capacity for everything spiky, hit 75 to 85% utilization on their permanent staff. The shops who try to staff the peak hit 55 to 65% and bleed cash on the bench every Q1.

For a deeper read on the mechanics, the engineering team as a service playbook breaks down how agencies structure the permanent vs booked split.

The seasonal slumps you can set your watch by

Three predictable revenue troughs hit dev agencies every year. The owners who survive them have already modeled them into the runway calculation.

• Mid-December through mid-January: clients defer approvals, AP teams go on holiday, new SOWs do not get signed. Expect 30 to 50% revenue compression for 4 weeks. Budget for it; do not let it surprise you.
• Mid-July through August: similar slowdown in EU markets especially. US tech is less affected but still soft. Budget for 15 to 25% compression.
• End of fiscal year for your largest enterprise clients: budgets freeze 30 days before close. If your client's FY ends in March, expect February to be tight.

Practical move: bill the December invoice on November 25 with net-15 terms so it clears before the holiday slowdown. Bill the January work on January 5 even if the client is not back yet, so the AP clock starts ticking. Move quarterly retainer renewals to October and April, not December and January.

The agencies who plan for the December dip in October keep their team intact through Q1. The ones who notice it on December 28 are the ones laying off engineers on January 15.

Invoice factoring: the lever to pull last

Invoice factoring is selling your outstanding invoices to a third party (factor) for 80 to 90% of face value up front, with the remaining 10 to 20% (minus a fee) paid when the client settles. Common in agency-land but expensive: effective annualized cost typically runs 18 to 40% APR.

When factoring makes sense:

• You have a single large net-60 enterprise invoice that is creating an acute payroll gap, and you cannot collect early.
• You are bridging to a known revenue event (a signed contract starts in 6 weeks).
• The factor's fee is lower than the cost of losing a key engineer.

When factoring is a trap:

• You are using it monthly to cover normal payroll. That is not bridging, that is borrowing at 30% APR forever.
• You have not first tried offering a 2/10 net-30 discount (2% off if paid within 10 days). Many clients take it and you save 25%+ in effective cost.
• Your client is small and the factor will treat collection aggressively, damaging the relationship.

Before you call a factor, exhaust the cheaper options: a small line of credit at your bank (usually 8 to 12% APR), the 2/10 discount, a payment-plan offer to the client, or in extremis a short founder loan to the business documented properly.

The cash-flow risk profile by client mix

The fastest diagnostic for an agency's cash-flow health is to look at the client-mix breakdown.

Client mix	Cash-flow risk	Typical terms	What kills you
80% founders / seed startups	Medium	Net-15 to net-30, deposits accepted	Client business failure, deal volatility
50/50 startups + mid-market	Low	Net-30 mostly	Concentration in one large client
30% startups, 70% mid-market SaaS	Low to medium	Net-30, occasional net-45	A slow AP cycle hitting at year-end
Heavy enterprise (60%+)	High	Net-45 to net-60	Procurement freezes, single-PO concentration
One anchor client at 50%+ of revenue	Critical	Anchor's terms dominate	Anchor leaves, anchor delays a payment

The healthiest agencies we see run roughly 40/40/20: founders, mid-market, one or two enterprise logos for credibility. No single client above 25% of monthly revenue. Mix of payment terms that average net-25 across the book.

If your mix is heavy enterprise, the lever is not "stop selling to them," it is "price the cash risk in." A 25% markup on net-60 work pays for the line of credit you will need to bridge those gaps, and still leaves margin.

The agency partner angle

Many agencies stabilize cash flow by adding two structural revenue paths that do not require new headcount.

First, white-label execution: take overflow work, ship under your brand, with delivery running through booked engineers from a marketplace. You bill at your agency rate ($150 to $300/hr); the booked engineer costs $12.50 to $50/hr equivalent (Cadence at $500 to $2,000/week ÷ 40). The margin covers PM time and goes straight to cash flow. The white-label development services breakdown covers how to set this up cleanly.

Second, the referral angle. Cadence pays partners 10% recurring on every founder you refer, for the lifetime of their account. For an agency talking to 20 founders a month, even a 10% conversion is meaningful passive revenue. Start with the partner program for a recurring 10% on every referral; the contract is straightforward.

These are not magic; they are structural. They show up on the P&L and smooth out the lumpy project-based cash flow that defines most agencies.

What to do this week

1. Pull your AR aging report today. Identify every invoice over 30 days late. Call (do not email) the AP contact for each one.
2. Calculate your runway. Cash on hand divided by last month's total burn. Write the number down.
3. If your runway is under 3 months, freeze hiring and any non-billable spend until you are back above 3.
4. Audit your last 5 SOWs. How many had a 50% deposit? If the answer is less than 5, fix your template.
5. Move any net-60 client to net-30 on renewal, or add a 20% rate premium. Walk away from clients who refuse both.

If you are running an agency and want a faster way to handle spiky overflow without taking the bench-cost hit, look at booking engineers by the week with no notice period through Cadence's partner channel. 48-hour free trial, weekly billing, replace any week, and you earn 10% recurring on any founder you refer.

For more on the invoicing side of all this, our dev agency invoicing best practices guide covers the five operational rules (weekly billing, itemized line items, auto-debit, late fees, and reminder cadence) that pair directly with the cash-flow playbook above.

FAQ

What is a healthy days-sales-outstanding (DSO) for a dev agency?

Under 35 days is healthy. 35 to 45 is workable. Above 45 means you have a collections process problem, not a sales problem, and you should hire a part-time bookkeeper before your next engineer.

Should I take net-60 enterprise clients?

Only if (a) they represent less than 30% of monthly revenue, (b) you have a written late-payment clause you will actually enforce, and (c) the rate is 20 to 30% higher than your standard net-30 rate to compensate for the cash float.

How much runway should a dev agency keep?

Three months minimum for survival. Six months for strategic freedom (the ability to walk away from bad clients, weather a soft quarter, or hire opportunistically). Most agencies operate at one month and discover the problem in January.

Is invoice factoring worth it?

Only as a bridge for a specific, known cash gap, not as recurring funding. The effective cost is 18 to 40% APR. Always try the cheaper options first: a 2/10 net-30 discount to the client, a small bank line of credit, or a direct payment-plan negotiation.

How do I price in the cost of slow-paying clients?

Add a 20 to 30% premium to your standard rate for net-60 terms, and a 10 to 15% premium for net-45. This funds the line of credit (or self-funded float) you need to bridge their slow payment cycle, so the cash drag does not erode margin.