Agencies
How to Hire a Social Media Agency in the US: A Buyer's Checklist for 2026
28 August 2026 · 15 min read · By Orion Media Group
Hiring a social media agency in the US is unusually easy to get wrong, because the sales process rewards the wrong signals. The agencies with the best decks are not reliably the agencies with the best editors, and the case studies you are shown are selected precisely because they are exceptional.
This is a buyer's checklist, written from the other side of the table. It covers how to scope the work before you talk to anyone, how to shortlist without wasting six weeks, how to run a paid pilot that actually predicts performance, which contract clauses matter in US agreements, and the red flags worth walking away from.
Use it whether you end up hiring us or someone else. A well-scoped brief makes every agency you talk to perform better, including the one you eventually choose.
Step 1: Decide which problem you are actually buying a solution to
Businesses tend to describe the symptom — "our social media isn't working" — rather than the constraint. There are only a handful of real constraints, and they lead to completely different hires.
- Production constraint: you know what to say but cannot produce enough finished content. Hire a production-led content agency.
- Distribution constraint: you produce content but it does not reach anyone. Hire for editing depth, hook craft and platform-native packaging.
- Conversion constraint: content reaches people but nothing happens commercially. Hire for funnel and offer work, not more posting.
- Consistency constraint: quality is fine but output collapses whenever the business gets busy. Hire for process and cadence guarantees.
- Attention constraint: nobody internally owns social at all. Hire full channel management, not just editing.
Write your constraint in one sentence at the top of your brief. Agencies that respond to that sentence directly are listening; agencies that respond with a generic capabilities deck are not.
Book a callStep 2: Set the budget band before the first call
Withholding budget in the hope of getting a cheaper proposal generally produces a worse one. Without a band, agencies default to scoping for their median client, and you spend three calls discovering the proposal was never aimed at your situation.
US market bands are reasonably consistent: under $1,500 per month buys scheduling and light management; $1,500–$3,000 buys genuine production volume on one to three platforms; $3,000–$8,000 buys multi-platform management with substantial video output and dedicated account leadership; $10,000 and above buys paid media and campaign production alongside organic.
State your band and ask each agency what they would cut and what they would prioritise at that number. The answer reveals their judgement more clearly than any case study.
Step 3: Shortlist to three, not thirteen
Long shortlists cost you weeks and produce worse decisions, because by the tenth call the proposals blur. Three is enough to triangulate pricing and approach.
Build the shortlist from evidence rather than directories: accounts in your industry whose content you genuinely stop scrolling for, referrals from peers who can show you the actual output, and agencies whose own social presence demonstrates the craft they are selling. That last filter removes a surprising number of candidates.
- Check their own channels. An agency selling short-form that posts nothing but static quote graphics is telling you something.
- Ask for three accounts grown from a low base, not three accounts that were already large when they arrived.
- Verify at least one reference by asking about turnaround and revision handling, not about whether they were nice to work with.
Step 4: Run a paid pilot instead of a longer sales process
The highest-signal step in the entire process is giving two shortlisted agencies the same raw footage and the same brief, paying each for one round of work, and comparing what comes back. A pilot costs a fraction of a bad twelve-month retainer and tells you more than any number of calls.
Judge the pilot on four things: speed to first cut, whether the hook holds a stranger's attention with the brand stripped out, how the captions and safe zones hold up watched on an actual phone rather than a laptop preview, and how the agency responds to your revision notes. Defensiveness at the revision stage during a pilot becomes friction every week thereafter.
Watch every pilot cut on your phone, sound off, while scrolling. That is the environment the content has to survive in.
Book a callStep 5: Read the contract clauses that actually matter
- Term and exit: three months is a fair minimum. Insist on a 30-day exit after the initial term and refuse silent auto-renewal.
- Ownership: you own raw project files and final masters. Agencies that retain project files are engineering switching costs.
- Scope definition: asset counts by format and platform, in numbers, not adjectives.
- Turnaround: named business-day windows for first cut and revisions, with a remedy if they slip repeatedly.
- Revisions: how many rounds are included and what triggers additional fees.
- Account access: you own the platform accounts, ad accounts and analytics; the agency gets delegated access that can be revoked instantly.
- Key-person clause: if you are hiring partly because of a specific editor or strategist, name them.
Step 6: Agree the scoreboard before month one
Agencies drift toward reporting whatever looks best. Fix the scoreboard before work starts, and keep it short enough that everyone remembers it.
For most US businesses three metrics are enough: a leading creative metric (average watch-through or three-second retention), an audience metric (follower velocity or qualified new followers), and a commercial metric (inbound conversations, booked calls, demo requests or attributable orders). Reach and impressions are context, not targets.
Set the review cadence at monthly with a 90-day checkpoint. At 90 days the honest question is not "did we go viral?" but "is the leading creative metric improving, and has the commercial metric started to move at all?" If neither has, the problem is either the offer or the agency, and it is worth naming which.
Red flags worth walking away from
- Guaranteed follower counts or guaranteed viral results — these are bought, not earned.
- No named editor, or an inability to say who touches your account week to week.
- Twelve-month lock-in with no exit clause.
- Reporting that leads with impressions and never mentions watch-through or inbound.
- Refusal to run a paid pilot on your actual footage.
- Case studies without dates, baselines or context — a percentage with no starting number means nothing.
- Pressure to sign before the end of the month at a discount that expires.
What good looks like 90 days in
A working agency relationship at the 90-day mark has a recognisable shape. Content arrives on schedule without chasing. The hook quality is visibly better in month three than in month one, because performance data has been fed back into production. You can name the two formats that consistently outperform for your audience, and so can the agency.
Commercially, you should be hearing the content mentioned — prospects referencing a video on a sales call, inbound messages that arrive already warm, or customers who describe having watched you for weeks before reaching out. Those qualitative signals usually appear before the dashboard moves, and they are the earliest reliable evidence that the engine is working.
Frequently asked questions
- How long should the first contract with a social media agency be?
- Three months is a reasonable minimum because content compounds and the first month is largely setup. Avoid twelve-month terms without a 30-day exit clause after the initial period.
- Should we run an RFP?
- For most US small and mid-market businesses, no. A one-page brief stating your constraint, budget band, required asset volume and success metric, sent to three shortlisted agencies, gets better answers faster than a formal RFP.
- What should a paid pilot cost?
- Typically a few hundred to a couple of thousand dollars for one round of real work on your own footage. Treat any refusal to pilot on your material as a data point.
- Who should own the social accounts?
- You, always. The agency should have delegated access through business-manager tools that you can revoke immediately. Never let an agency create your accounts under their own ownership.
- What if the agency is not working after 90 days?
- Ask whether the leading creative metric improved. If watch-through and retention rose but commercial results did not, the constraint is likely your offer or funnel. If the creative metric is flat, the agency is not iterating and it is time to exit.