Most SDR agencies quote a meeting count and call it a result. Here are the five questions- denominator, headcount, window, definition, and segment, that reveal what that number actually means, plus how to calculate the real cost per meeting.
Sachin Shah
September 21, 2026
6 min read
Apply every question in this post to us. If we cannot answer them, the argument applies to us too.
How to Evaluate an SDR Agency: 5 Questions That Expose Their Real Numbers -
Every SDR agency you talk to will tell you their BDRs book meetings. Almost none of them will tell you what it took.
"We booked 50 meetings for a client last quarter" is not a result. It is a headline with the denominator removed. Fifty meetings from three BDRs in two months is a machine. Fifty meetings from twelve BDRs in six months is the same headline describing a business that is twelve times worse at this. Per BDR per month, the first is roughly eight meetings and the second is under one.
"
"Same sentence. Same number. Two completely different companies."
You do not need industry benchmark data to catch this. You need five questions, and the arithmetic is something you can do in your head while they are still talking.
1
1.
How many dials and emails did that take?
2
2.
How many BDRs produced it, and what did each meeting cost?
3
3.
Over what window - ramp or steady state?
The first three questions help you understand the activity, team effort, and time period behind the reported meeting numbers. But to understand what those numbers actually represent, you also need to look at how a meeting was defined and who those meetings were with.
1
4.
What counted as a "meeting"?
2
5.
What ACV and what persona?
How many dials and emails did that take?
This is the denominator, and it is the one agencies are least prepared for. Not because the number is embarrassing, but because most of them do not track it at the pod level and have never been asked.
Watch what happens when you ask. An agency that runs a real operation gives you a number, possibly a rough one, immediately. One that does not will reframe: "we focus on quality over volume," or "it varies by client." Both may be true. Neither is an answer, and the reframe tells you the reporting does not exist.
If you want to see what a published denominator looks like, ours is in how we review a BDR call: dials, connects, conversations past 90 seconds, and meetings, per BDR per week.
How many BDRs produced it, and what did each meeting cost?
Meetings scale with headcount. Any number scales with headcount. Without the caller count you cannot tell a productive pod from an expensive one, and agencies know this, which is why case studies quote totals rather than per-head figures.
Ask for meetings per BDR per month. Then do the only calculation that matters commercially: cost per meeting. Take the monthly retainer, divide by meetings per month, and compare it against your average deal size and close rate off a booked meeting. If cost per meeting is a meaningful fraction of gross profit per deal, the model does not work no matter how good the agency is.
Make them do this with you on the call. An agency that has never calculated cost per meeting for a prospect is optimising for meetings rather than for your economics.
Over what window, and was that ramp or steady state?
Every outbound engagement has an ugly middle. Weeks four to six are when connect rates sit below expectation, the coach is still fixing openers, and the client starts asking pointed questions. Numbers pulled from months one and two describe a team learning. Numbers from months four onward describe a team working. We wrote about that ramp in more detail in how we build US pipeline with a team in India.
An agency quoting you a six-month total has averaged the ramp into the result, which understates their steady state. One quoting you their best month has done the opposite. Neither is dishonest. You just need to know which one you are holding.
The follow-up that matters more: how long until steady state? If the answer is under four weeks, they are either inheriting a warm pipeline or telling you what you want to hear.
What counted as a meeting?
This is the question almost nobody asks and the one with the widest spread.
"Meeting" can mean booked on the calendar, held, held and accepted by your AE, or converted to a qualified opportunity. Those are four different numbers and the gaps between them are large. Cold-booked meetings with senior buyers no-show at meaningful rates, so booked and held can differ by a third. Held and accepted differ by however strict your AEs are.
Definition
What it means
Why it's inflated vs. reality
Booked
Placed on the calendar
No-shows aren't subtracted - can overstate real output by a third
Held
Prospect actually showed up
The honest baseline; use this in your contract
Held & Accepted
Showed up, and your AE deemed it qualified
Depends entirely on how strict your AEs are
Converted to Opportunity
Progressed into a real qualified deal
Furthest from the marketing number, closest to revenue
An agency quoting booked meetings against a competitor quoting held meetings is not a comparison. Get everyone to the same definition before you compare anything, and make sure the definition in the contract is the one you actually want, which is almost always held rather than booked. If they are paid per booked meeting, you have just created an incentive to book people who will not show.
What ACV and what persona?
This is the question that decides whether any of the above transfers to you.
A $5,000 self-serve tool sold to a marketing manager and a $60,000 platform sold to a VP of Engineering do not share a playbook. They do not share a sequence length, a connect expectation, or a definition of a good reply. The manager answers their own phone. The VP has a calendar defended by two other people and screens hard.
As deal size and seniority go up, connect rates fall. That is not a performance problem, it is the buyer. What changes is where the work goes: fewer, better-researched touches instead of volume, and a longer sequence because the buying committee is bigger and slower. An agency with an impressive meetings-per-BDR number selling $8,000 deals to managers tells you nothing about whether they can reach your VP of Engineering.
So the full version of question five is: show me those numbers from an engagement at my ACV band, calling my persona. If they cannot, they may still be the right partner. But you are now buying on judgment rather than evidence, and you should know that is what you are doing.
Turning this on us -
We publish our enterprise funnel: dials, connects, conversations past 90 seconds, and meetings booked, per BDR per week, India calling into the US. It is not a flattering number in isolation and we publish it anyway, because the denominator is the only thing that makes the numerator mean anything.
What we do not publish is a cross-client benchmark table broken out by ACV band and persona. We could write one. We would be pooling too few engagements per cell for it to be worth anything, and publishing thin data with confident formatting is the same trick we are describing here, dressed better. When the sample supports it, we will publish it.
Where this checklist misleads:
Low meetings per BDR is not automatically bad. In enterprise it is expected, and an agency whose enterprise numbers look like mid-market numbers is either counting something loosely or not actually reaching the buyer they claim. High dial volume is not automatically good either - it often means the list is broad, which is a list problem wearing an activity costume. And none of these five questions tells you whether the agency can sell your product specifically. They tell you whether it runs a measurable operation. That is a necessary condition, not a sufficient one. The rest you learn from how they handle discovery, and specifically from whether they tell you which parts of your motion they think will not work.
The point :
You do not need a benchmark report to evaluate an outsourced SDR team. You need the denominator, the headcount, the window, the definition, and the segment. Five questions, two minutes, no data required on your side.
If an agency cannot produce them next to their meeting count, you are not looking at a case study. You are looking at a highlight reel.
"Frequently Asked Questions"
There is no universal number, and any agency quoting one without knowing your ACV is guessing. Calculate it yourself: monthly retainer divided by meetings held per month. Then check it against gross profit per closed deal and your close rate off a booked meeting. Enterprise motions carry a far higher cost per meeting than mid-market ones, and that is expected rather than a red flag.
It works when you have a defined ICP, a repeatable motion, and an ACV high enough that one closed deal covers many months of acquisition cost. It fails when sales has been entirely founder-led with nothing written down, because there is no motion to systematise yet, and when the buyer is undefined and the plan is to start calling and find out.
First meetings typically appear during ramp, but a stable, predictable rate takes months, not weeks. Weeks four to six are the hardest part of any engagement. The most common reason these engagements fail is a client pulling the plug in week five, which is usually right before the curve turns.
Define "meeting" as held rather than booked, and make sure the payment trigger matches that definition. Also agree upfront on what reporting you receive weekly: dials, connects, conversations, meetings, and the list segments they came from.
Get a second opinion on your agency's numbers
Send us the numbers an SDR agency has quoted you. We'll tell you what the gaps are - including when they look fine.
Sachin Shah