In-House SDR vs Outbound Agency: Cost Comparison and ROI Model

Short answer

There is no universal winner. The right answer depends on what your company can already do, what it needs to learn, and how quickly it needs to learn it.

• An in-house SDR usually makes sense when you have real capacity to recruit, train and manage the role, when product knowledge is difficult to externalise, and when you want to build a permanent commercial function.
• An agency can make sense when the offer and the ICP already have some level of validation, but you lack infrastructure, operating experience or the speed to launch and learn.
• A hybrid model can work when the internal team keeps knowledge, strategy and closing, while an external partner operates data, infrastructure, campaigns or conversation generation.
• Not activating outbound yet is the correct decision when the offer is still weak, the ICP is uncertain, the margin cannot support the cost of acquisition, the addressable market is too small, or nobody can attend and convert the meetings.

The comparison that matters is not salary versus fee. It is total cost of the system, time to learning, and cost per qualified meeting.

Why the usual comparison is set up wrong

The common version of this decision puts a base salary on one side and a monthly retainer on the other. Those are not comparable line items. A base salary excludes employer contributions, recruitment, onboarding, tooling, data, sending infrastructure, the manager's time and the productivity lost during ramp-up. A retainer excludes internal coordination time, the tools you still pay for, and the fact that message review and commercial follow-up remain your responsibility.

The size of that gap is not a matter of opinion. In the United States, the Bureau of Labor Statistics reports that for private industry workers in March 2026, wages and salaries accounted for 69.9% of total employer compensation costs and benefits for the remaining 30.1% (BLS, Employer Costs for Employee Compensation, released 12 June 2026, USD, consulted 27 July 2026). In the European Union, Eurostat estimates that non-wage costs represented 24.8% of total labour costs in the EU and 25.6% in the euro area in 2025, ranging from 4.8% in Romania to 32.3% in France (Eurostat, 31 March 2026, EUR, consulted 27 July 2026). These are two different official measures on two different bases; they should not be averaged into a single global figure, and neither is a substitute for your own payroll numbers.

So the correct unit of comparison is not a single number. It is a set of them.

• Total monthly cost, including everything amortised.
• Time until the operation actually functions.
• Cost per qualified meeting, using your own definition of qualified.
• Learning generated — what you know at the end of the quarter that you did not know at the start.
• Risk carried, and by whom.
• Internal capacity required, whichever route you take.
• Expected gross value, not nominal pipeline.

Total cost of an in-house SDR

An internal SDR is a system, not a salary. The components below are the ones that change the answer most often. Their size depends heavily on your country, contract type and seniority, so treat the list as a structure to fill in rather than a set of numbers to copy.

• Fixed salary.
• Variable compensation or commission.
• Employer contributions and benefits — the official share of total labour cost varies widely by country, as the BLS and Eurostat figures above show.
• Recruitment: advertising, agency fees if used, and the hiring team's time.
• Onboarding: product training, market context, messaging and tooling.
• Lost productivity during ramp-up, before the role reaches normal output.
• Management time: coaching, call reviews, pipeline reviews and quality control.
• Tools: CRM seats, sequencing platform, dialler, meeting scheduling.
• Data and enrichment.
• Domains, mailboxes, warm-up and sending infrastructure.
• Ongoing training and quality control of the messaging.
• Turnover: the probability that the role is vacant again within the year.
• Amortised costs: recruitment, onboarding and turnover spread across expected tenure rather than charged to a single month.
• Opportunity cost: what the manager and the founder would otherwise be doing with those hours.

Two cautions. First, do not convert a salary or a contribution percentage from one country into a global reference — labour cost structures are not transferable. Second, ramp-up is a cost even when nothing goes wrong: the role is paid in full while output is still building, and that gap belongs in the model.

Total cost of an outbound agency

An agency engagement is also a system, and its published price is only part of it. Provider pricing pages are useful for understanding commercial models — retainer, setup plus retainer, retainer plus performance — but a price published by a supplier is a commercial claim, not neutral market data. The same applies to tool costs: check the vendor's own pricing documentation at the time you build your model, because per-seat and per-credit pricing changes frequently.

• Setup or onboarding fee, usually one-off.
• Monthly fee.
• Tools or data not included in the fee.
• Performance-based variable, where one exists.
• Internal coordination time: briefings, reviews, approvals, weekly calls.
• Onboarding the provider into your product, market and objections.
• Knowledge transfer, in both directions.
• External dependency: what happens to the operation if the relationship ends.
• Risk of low-quality output reaching your market under your name.
• Switching costs if you change provider: domains, data, sequences, learning.
• The internal involvement that will still be required regardless.

Outsourcing execution does not outsource ownership. ICP definition, message review, objection handling and commercial follow-up stay with you.

This is the most common reason an agency engagement underperforms. The company treats it as a purchase rather than a partnership, stops feeding it market knowledge, and then evaluates the result on meeting count alone. If nobody internally can answer a technical objection, review the message against how the market actually buys, or run the meeting well, no supplier can compensate for that.

Comparison beyond cost

Cost is one axis. The table below compares the two structures on the dimensions that usually decide the outcome. Neither column is the winner; each row names the context that favours one or the other.

Decision model by context

Four possible decisions, and the conditions that favour each. Read them as thresholds, not preferences.

The fourth option is the one most often skipped. If the constraint is upstream of access, adding prospecting multiplies the cost of an unresolved problem — which is the argument developed in where outbound actually fits.

Comparative calculator

Fill in the two columns with your own figures. The calculator runs entirely in your browser: nothing is stored, nothing is transmitted, and no personal data is requested. If you want to see the mechanics before entering real numbers, load the illustrative example — those values are round and neutral placeholders, not a market benchmark.

One warning about interpretation. Meetings booked are not meetings held; meetings are not opportunities; opportunities are not sales; pipeline is not revenue; and revenue is not gross margin. The model deliberately works in expected gross margin so the output cannot be mistaken for booked revenue.

Decision examples

The following four scenarios are hypothetical illustrations of the decision logic. They are not clients, and they contain no client data.

1. 1Hypothetical: a SaaS company with a proven offer in its home market wants to test a second country. It has validation and margin but no local list, no local infrastructure and no local reference. An agency or a hybrid engagement lets it test the market before committing to a local hire — the point of the test is learning, not headcount.
2. 2Hypothetical: a company with a complex technical sale, where the first conversation already requires engineering-level understanding. Product knowledge is the constraint, not access. An internal hire — or a hybrid where the external partner generates access and an internal specialist runs every conversation — fits better than fully externalised prospecting.
3. 3Hypothetical: an existing sales team is closing well but has no capacity left to prospect. The constraint is hours, not knowledge. A hybrid model that adds top-of-funnel capacity without duplicating management structure is usually the least disruptive route.
4. 4Hypothetical: a company without a clear ICP hires more prospecting in the hope that volume will reveal the pattern. Neither an SDR nor an agency solves this. The right decision is to not activate outbound yet and resolve the offer and the ICP first.

Questions to answer before deciding

• Does the offer already generate sales?
• Is there an identifiable ICP?
• What is the gross margin per deal?
• How much time can a manager genuinely dedicate?
• Who will handle the replies?
• Who will run and follow up the meetings?
• How much product knowledge does the first conversation require?
• What learning speed does the company need?
• Which part of the system must stay internal?
• How will a qualified meeting be defined, in writing, before you start?
• What happens if the channel does not work in the first months?

If several of these answers are unclear, the constraint may not be outbound capacity at all. The growth bottleneck audit covers how to test that before committing budget either way.

Methodology and limitations

What the model includes: direct compensation and employer costs, recruitment, onboarding and ramp-up, tools, data, sending infrastructure, management time and turnover risk on the in-house side; setup, fee, variable, tools and data, internal coordination and other attributable costs on the agency side. Recruitment, setup, onboarding and turnover are entered already amortised — divide the one-off cost by the number of months you expect the arrangement to last.

• Employer contributions, benefits and dismissal costs are country-specific. The official BLS and Eurostat measures cited above describe different bases and cannot be combined into one global average.
• The model does not automatically include every tax or employment cost in your jurisdiction. Check your own payroll figures.
• Results change with the offer, the ICP, the market, timing, ticket size, margin and commercial execution.
• A qualified meeting does not guarantee a sale, and the definition of qualified is yours, not an industry constant.
• The output is determined entirely by the inputs. It is an estimate produced from your assumptions, not a forecast.

Figures deliberately omitted: average SDR salaries by country, average agency retainers, average ramp-up length and average turnover rates. Public numbers for these exist, but the reliable ones are either country-specific or come from vendor surveys with a commercial interest in the answer. Rather than publish a composite average that would not survive contact with your own payroll, the calculator asks for your figures and explains what belongs in each field.

Where Cardo fits

For transparency: Cardo operates B2B outbound systems, which is a commercial interest in this topic. That is precisely why this guide models four outcomes rather than one, and why it does not claim that an external partner is cheaper or produces a better return. In several of the contexts above, hiring internally or waiting is the better decision.

See how Cardo designs and operates B2B outbound systems, or read why a specialist partner model differs from a generalist agency before deciding which structure fits your case.