Every Series A founder we talk to has run this comparison in a spreadsheet, and most of them have run it with the wrong number for the SDR. Here is the arithmetic with the real one, and the part of the decision that price does not settle.
The SDR figures below are in US dollars, because that is the currency The Bridge Group reports and re-denominating someone else’s survey would invent a number nobody measured. Agency retainers are compared at par rather than at a rate that will have moved by the time you read this.
What does one SDR actually cost?#
The advertised number is on-target earnings. The Bridge Group's 2025 SDR Models, Metrics and Compensation report, drawn from 351 B2B companies with a median revenue of $47M, puts median SDR OTE at $80,000, split $55,000 base and $25,000 variable.
That is not the cost of employing one. Four things sit on top:
| Line | Annual | Where it comes from |
|---|---|---|
| On-target earnings | $80,000 | Bridge Group 2025 median |
| Employer taxes, benefits, equipment | ~$20,000 | Conventional 25% loading on OTE |
| Tooling and data | $6,000 to $12,000 | Sales Navigator, a sequencer, an enrichment budget |
| Share of a front-line manager | ~$23,000 | $146,000 manager OTE across 6.4 SDRs per leader |
| Loaded total | ~$130,000 |
The manager line is the one that gets left out. The same report puts SDR manager OTE at $146,000 and the rep-to-leader ratio at 6.4, so every SDR carries roughly $23,000 of management with them. One SDR does not need a sixth of a manager, they need considerably more, because a single rep with nobody experienced watching their sequences is the most common version of this hire failing.
How does that compare over three years?#
Show the numbers behind this chart
| Cumulative spend | One SDR, loaded |
|---|---|
| Year 1 | $130,000 |
| Year 2 | $260,000 |
| Year 3 | $390,000 |
A mid-market agency retainer is typically closer to half a loaded SDR over a full year, and the gap is wider than it looks because the two lines do not start producing at the same time.
What does the ramp do to year one?#
Average ramp is 3.0 months, the lowest The Bridge Group has recorded since 2010. Average tenure is 1.9 years. Put those together and each hire gives you about 20 productive months, against 23 months of salary.
Median annual attrition across the 351 companies surveyed is 40%. That is not a number about bad hiring; it is the structural churn rate of the role, and it means the ramp clock restarts more often than a three-year plan assumes. Every restart costs three months of output, a recruiting cycle, and the institutional memory of what the last person learned about which segments reply.
A system has a ramp too. Ours is three months, which is the same. The difference is that the system's ramp happens once.
Which one actually fits your motion?#
Price is the easy part of this decision and usually the wrong part to decide on. The real question is whether your pipeline is limited by consistency or by conversation.
| Hire the SDR | Build the system | |
|---|---|---|
| Works when | A small named account list, high ACV, long cycles, relationship-led | A large addressable market, repeatable qualification, €10K+ ACV |
| Fails when | The list is too large to work by hand | The buying decision needs a person from the first touch |
| Time to first output | ~3 months of ramp | ~3 months of build |
| What breaks it | Attrition, and the knowledge leaving with them | Data decay and stale sequences, if nobody is watching |
| What you own after | The relationships that person built | The targeting, the logic, and the documentation |
The pattern we see most often at Series A is a founder who is still the best closer in the company, a market large enough that hand-working it is not possible, and a list nobody has ever properly defined. That combination is a system problem wearing a headcount costume, and hiring one SDR into it produces an expensive researcher.
Why does hiring one SDR so often conclude that outbound does not work?#
Because one is a sample size, not an experiment.
When a single rep's numbers come in flat, there is no way to tell whether the segment was wrong, the message was wrong, the data was wrong, or the person was having a hard quarter. All four look identical from the outside, so the team picks the cheapest explanation, which is usually that outbound is dead in their market.
A system fails more legibly. If enrichment coverage is 40% on one segment and 85% on another, that is visible. If replies come from one persona and not the other, that is a number rather than a hunch. It is not that a system is smarter than a good SDR. It is that a system produces enough comparable observations to tell you which thing to fix.
What does the end state look like?#
Both, in a specific order.
The version that works is a system doing targeting, enrichment, sequencing and routing, and a person spending their day on the conversations it surfaces rather than on building lists in a spreadsheet. Reps who trust the leads they are handed work them properly, and reps who do not, do not.
The order matters. Building the system first and hiring into it means the SDR's first day is spent on live conversations. Hiring first and asking that person to build the system means paying $130,000 a year for someone to learn Clay while their sequences sit idle, and then watching that knowledge walk out at the 20-month mark.
Questions we get asked about this
- How much does an SDR really cost per year?
- The Bridge Group's 2025 report puts median SDR on-target earnings at $80,000, split $55,000 base and $25,000 variable. That is the advertised number. Add roughly 25% for employer taxes, benefits and equipment, add $6,000 to $12,000 of tooling, and add a share of the manager the role needs, and the loaded figure lands near $130,000 a year in the US. In Germany the base is lower and the employer loading is higher, so the total is closer than founders expect.
- How long before a new SDR is productive?
- Average ramp is 3.0 months according to The Bridge Group's 2025 research, the lowest since 2010. Combined with average tenure of 1.9 years, that means you get roughly 20 productive months per hire. Median annual attrition of 40% means the ramp clock restarts more often than most hiring plans assume.
- Is one SDR enough to build a pipeline motion?
- Usually not, and this is the trap. One SDR has no comparison group, so when a segment does not respond you cannot tell whether the segment is wrong, the message is wrong, or the person is having a bad month. Teams that hire one SDR to test outbound often end up concluding outbound does not work, when what they actually ran was a sample size of one.
- When should we hire the SDR instead?
- When the motion depends on a human working a small number of accounts deeply rather than a system working a large number consistently. High ACV, long cycles, relationship-led selling into a named account list, or a market small enough that every target deserves a hand-written approach. In those cases a system adds overhead and the person adds the thing that actually closes.
- Can we do both?
- That is the common end state and the one we build towards. The system handles targeting, enrichment, sequencing and routing, and the SDR spends their day on conversations rather than on list building. The order matters though: hiring the person first and asking them to build the system is how you get an expensive researcher who leaves in eighteen months.
- What happens to the system if we bring outbound in-house?
- It transfers. The tooling accounts are in your name, the logic is documented, and the handover happens whenever you ask for it. A system you cannot run without the agency that built it is a system built badly, and it is the reason we quote a three-month build rather than an indefinite retainer.
Co-Founder of Harochi, a Berlin-based GTM engineering agency. Previously at Google in New York, then building outbound systems at Leapsome and UPPER. Gets called in when a Clay or CRM build has already gone wrong.
Connect on LinkedInTell us where pipeline breaks.
30 minutes with Macklin. We will tell you what we would build first, and whether it is worth paying us to build it.