An outsourced SDR agency buys you speed and removes hiring risk; an in-house sales development team buys you control, learning, and lower unit cost after roughly month eight. Agencies typically run $5,000 to $15,000 per month per SDR pod, against $95,000 to $130,000 fully loaded for one in-house SDR. Neither fixes a weak offer or an undefined market — agencies fail most often because the buyer had no qualified target list and no internal owner, not because the agency could not send email.
This guide is about outsourced sales development — hiring an external agency to run outbound prospecting and book qualified meetings for your sales team. If you arrived looking for software-defined radio, this is the wrong page. SDR here means sales development representative.
The decision usually arrives under pressure. A revenue target moves up, the founder can no longer prospect personally, and hiring an SDR means eight weeks to fill and three months to productivity. An agency promises meetings in three weeks. That trade is real, but it is priced and structured in ways that make it easy to sign something you cannot judge.
Below: what an agency actually delivers, a decision matrix against the two alternatives, a six-month cost model with real numbers, the contract clauses that cause disputes, and a 90-day pilot design that produces a clear keep-or-kill answer. For the underlying role definitions, see SDR vs BDR and what an SDR does.
What an SDR agency actually delivers — and what it does not
Agencies bundle five things. Knowing which you are buying prevents the most common disappointment, which is expecting strategic market work from a team priced for execution.
- Trained headcount, immediately. People who have run outbound before, with a manager, available in two to three weeks. This is the core of what you are paying for.
- Infrastructure and deliverability. Sequencers, dialers, warmed sending domains, and inbox rotation. Genuinely valuable, because a mishandled domain warm-up can damage your primary email reputation for months.
- A data layer of variable quality. Most agencies resell a standard B2B database. This is where quality diverges most sharply between vendors, and where you have the most leverage to insist on your own list.
- Messaging execution. They will write and iterate sequences. They will not discover your positioning — if you cannot articulate why a buyer switches, no amount of copy testing produces it.
- Reporting on activity, and sometimes on outcomes. Volume, opens, replies, meetings booked. Whether meetings held and pipeline created appear in the report is a contract question, not a capability question.
What they do not deliver: your ICP, your qualification criteria, your follow-up discipline, or the institutional learning from a thousand rejections. That last one is the strongest argument for in-house, and it is why agencies work best as a bridge or a parallel channel rather than a permanent replacement for a sales development function.
Agencies compensated on meetings booked will book meetings. If qualification is loose, your reps spend their week on calls with people who cannot buy, and the contract still shows a green number. Define qualification before signing — title band, company size, budget authority, and a stated problem — and pay against meetings held and accepted by your rep, not scheduled.
Agency vs in-house vs AI-assisted prospecting
There are three ways to staff outbound, and most teams end up combining two. Compare them on the dimensions that change the answer rather than on a feature list.
| Dimension | SDR agency | In-house SDR | AI-assisted prospecting |
|---|---|---|---|
| Time to first meeting | 2–4 weeks | 3–5 months | Days |
| Monthly cost | $5,000–$15,000 per pod | $8,000–$11,000 fully loaded | Hundreds to low thousands |
| Commitment | 3–12 month contract | Employment, plus severance risk | Monthly, cancellable |
| Product depth | Low to moderate | High over time | Depends on your own team |
| Learning retained | Leaves with the agency | Stays in the company | Stays, if someone reviews it |
| Brand risk | Moderate to high | Low | Low if a human approves sends |
| Scales by | Adding pods | Hiring and training | Research volume, not headcount |
The pattern that works for most companies under $10M in revenue is not a choice between columns. It is AI-assisted research plus one in-house owner, with an agency added only when you need volume faster than you can hire and you already know the message converts. Buying agency volume before you know the message converts is how companies spend $60,000 learning their positioning was wrong.
The real six-month cost of each option
Monthly rates hide the comparison. The table below models six months from a standing start, including the costs buyers routinely leave out: ramp time, management attention, tooling, and the data layer.
| Line item | Agency (1 pod) | In-house (1 SDR) | AI-assisted (1 owner, part time) |
|---|---|---|---|
| Direct cost, 6 months | $48,000 (at $8,000/mo) | $47,500 (salary + benefits) | $3,000 (tooling) |
| Setup or recruiting | $2,500 onboarding | $12,000 (agency fee or time) | — |
| Tooling and data | Usually included | $6,000 (sequencer, dialer, data) | Included above |
| Your management time | ~3 hrs/week | ~6 hrs/week for 3 months | ~6 hrs/week |
| Productive months of 6 | ~5 | ~3 | ~5.5 |
| Six-month total | ~$50,500 | ~$65,500 | ~$3,000 plus your time |
Two conclusions follow. First, over six months the agency is genuinely cheaper than a first in-house hire, mostly because you skip recruiting and ramp — the speed argument is real, not marketing. Second, the crossover arrives around month eight to ten: the in-house SDR is now fully productive at a lower monthly cost and the learning stays with you, while the agency invoice does not decline.
The third column is not a like-for-like replacement, and treating it as one is a mistake in the other direction. It removes the research and list-building work — which is most of an SDR's hours — but somebody in your company still has to send, call, handle objections, and book. What it changes is how much outbound one person can credibly run, which is why it belongs in the picture regardless of which staffing model you pick. See AI BDR and B2B lead generation.
Both staffing models fail on the same input: a weak target list. Lessie searches 100+ live sources to find your accounts and the decision-makers inside them, with verified emails— so the agency or the rep starts from research, not a purchased CSV.
How agency pricing works, and the clauses that hurt
Four pricing models are common. Each shifts risk differently, and the one an agency pushes tells you what they are confident about.
- Retainer per pod. A fixed monthly fee for an SDR plus partial manager and tooling. Predictable, and the agency bears no outcome risk. Most common above $8,000 a month.
- Pay per qualified meeting. Typically $300 to $800 per meeting. Attractive because it looks outcome-based, but it makes the definition of “qualified” the entire contract. Expect pressure to loosen it.
- Retainer plus performance bonus. The most balanced structure in practice: a lower base with a bonus on meetings held or pipeline created. Prefer this if you can get it.
- Pilot then retainer. A discounted 60 to 90 day pilot. Read the auto-renew terms carefully — the discount often converts into a twelve-month commitment on a date you did not diarise.
Five clauses cause most disputes. Fix them before signature, because none of them are negotiable afterwards.
- Who owns the data. Contacts, sequences, replies, and call recordings should be exportable by you at any time, in a standard format. Assume that if it is not written down, you leave with nothing.
- Sending domain. Insist they use dedicated subdomains, never your primary domain. A burned primary domain outlives the contract by many months.
- Qualification definition and dispute process. Written criteria plus a mechanism to reject a meeting that did not meet them, with a credit rather than an argument.
- Named team and substitution. Who is actually working your account, and what happens when they are reassigned. Silent substitution is the most common quality collapse in month four.
- Exit and notice. Thirty days is reasonable; ninety with auto-renew is a trap. Diarise the notice date the day you sign.
A 90-day pilot you can actually judge
Most pilots end in ambiguity because too many variables moved at once. Constrain it so the result means something.
- 1Fix the market before the kick-off call
Supply the target list yourself, or approve theirs account by account. Define the segment narrowly — one industry, one size band, one region, one buyer role. A pilot across four segments produces four samples too small to read, and it lets both sides explain away the result.
- 2Write the qualification definition down
Four criteria, unambiguous: title band, company size, a stated problem in your category, and agreement to a next step. Then add the rejection path — your rep can mark a meeting unqualified within 48 hours and it does not count. Without this, the pilot measures scheduling, not pipeline.
- 3Name one internal owner with three hours a week
Somebody in your company must review messaging, listen to two calls a week, and answer product questions within a day. Agencies that fail almost always describe the same cause: no responsive counterpart. If nobody has the three hours, do not start the pilot.
- 4Instrument five metrics, not fifteen
Meetings held and accepted, qualified rate, opportunities created, cost per accepted meeting, and reply sentiment. Ignore opens and sequence volume entirely — they are the agency's inputs, not your outcomes, and reporting on them invites activity theatre.
- 5Set the decision rule in advance
Write down now what result means continue, what means renegotiate, and what means stop. A reasonable starting bar for mid-market B2B is 8 to 12 accepted meetings per pod per month at a cost per accepted meeting below one tenth of your average deal size. Deciding the bar after seeing the data is how bad contracts get renewed.
Run one control alongside it if you can: your own owner doing 10 researched, personalised touches a week using the same list. It costs almost nothing and it frequently reveals that the constraint was never headcount.
Where Lessie fits: the research layer both models need
Lessie is not an SDR agency and does not send email or book meetings on your behalf. It addresses the input that determines whether either staffing model works: knowing exactly who to contact and why now.
- Build the list yourself. Describe the account and buyer profile in plain language and get results from 100+ live sources, rather than accepting a resold database your competitors also rent.
- Verified contact details. Emails validated before export, which protects deliverability — the single largest hidden cost of outsourced outbound.
- A specific reason to reach out. Recent hiring, funding, product, and public activity give the first line something citable, which is what separates a researched touch from a templated one. See buying signals.
- Keeps the asset when the contract ends. The list and the research live in your account, not the agency's. See sales prospecting tools and pricing.
The honest recommendation: if you cannot yet describe who buys and why they switch, no agency will discover it for you at $8,000 a month — spend six weeks and a few hundred dollars proving the message with researched outreach first. If you already know it and simply need volume faster than you can hire, an agency is a rational bridge. Just diarise the notice date, and plan the in-house transition for around month nine.