Agencies are one of the most misused line items in an early-stage budget. Founders hire them too early, brief them too vaguely, manage them too loosely, and stay too long. Done well, an agency compresses six months of hiring into a two-week ramp and gives you access to a caliber of talent you cannot afford full-time. Done badly, you pay $25,000 a month for slide decks nobody reads.
Use an agency when you need a specific, bounded output that requires a skill you do not have in-house and will not need permanently. Rebranding, a paid-media launch, a category-defining PR moment, a compliance audit, a one-time video, a market-entry study.
Do not use an agency for anything that requires deep, ongoing context about your product, customers, or roadmap. That work belongs in-house even if it is slower and more expensive on paper. The hidden cost of an agency is the context you have to rebuild every quarter.
Contractor (1 person, 3–12 months): best when you know exactly what you want and need one pair of hands. Cheapest, highest ownership, lowest bench depth.
Agency (team, retained or project): best when you need multiple disciplines coordinated — strategist, designer, engineer, PM — and you need them to start Monday.
Hire: best when the work is permanent, strategic, and central to your moat.
If you are not sure, start with a contractor for four weeks. If the work is too big for one person, upgrade to an agency. If it never ends, hire.
Ask five founders you trust for referrals — never Google. Shortlist three. Give each the same paid discovery sprint ($5–15k, two weeks) with the same brief and the same deliverable. You will learn more from how they run the sprint than from any pitch deck.
Red flags in the sales process: senior partners on the pitch who disappear after signing, case studies that name-drop logos without describing outcomes, refusal to name the specific humans who will do your work, pricing that is a flat retainer with no scope attached.
A bad brief guarantees a bad output no matter how good the agency is. A good brief is one page and answers:
1. What is the business decision this work informs? 2. Who is the audience and what do we want them to do differently? 3. What does success look like in numbers? 4. What is explicitly out of scope? 5. Who owns the decision on our side, and by when? 6. What existing assets, data, and constraints must be respected?
If you cannot write this in one page, you are not ready to hire an agency. Spend another week internally.
Scope in an exhibit, not the SOW body — so you can amend without renegotiating the master agreement.
Kill fee of 30–50% for early termination, not 100%. If they insist on 100%, walk.
IP assignment on delivery and payment, with a carve-out for their pre-existing tools.
Named team in the SOW with a substitution-notice clause. This prevents the classic bait-and-switch.
Monthly, not quarterly, billing with a 15-day out. Retainers longer than 90 days without a break clause are traps.
Data and security addendum if they touch customer data. Non-negotiable.
Weekly 30-minute status call with a written agenda sent 24 hours before.
Weekly written update from the agency: what shipped, what is blocked, what changed, what they need from you.
Monthly retrospective — 45 minutes, both sides bring one thing that is working and one thing that is not.
Quarterly business review — is the work still tied to a business outcome, or has it become process for its own sake?
Assign one owner on your side. Agencies fail when the founder is the day-to-day contact for the first month, then hands off to a junior with no context in month two.
The weekly update becomes a status of activities instead of outcomes.
Deliverables slip and the excuse is always something on your side.
The senior person from the pitch has not been on a call in six weeks.
You find yourself re-briefing every meeting because nothing landed the first time.
Fire fast. Sunk cost is a tax on future decisions. Give 30 days written notice, pay the kill fee, get the IP, and move on. Do not try to fix a broken agency relationship — you will spend more managing the fix than you saved on the switch.
The best agencies work themselves out of a job. When you hire the in-house lead, the agency should transition the work in 60–90 days, document everything, and take a smaller advisory retainer or leave cleanly. Agencies that fight this transition are protecting revenue, not your business. That is the moment you learn which agencies to hire again and which to warn other founders about.
Used surgically, agencies are one of the highest-leverage moves in the early-stage playbook. Used lazily, they are the most expensive way to feel busy.
