SDR Metrics: Which Numbers the Rep Controls, and Which Measure the List
A sales development scorecard mixes three kinds of number and only one is about the person being reviewed. Sorting them by who controls what turns it into a diagnosis.
SDR metrics fall into three tiers: what the rep controls, what the list controls, and what infrastructure and the market control. Only the first belongs in a performance review. Accounts worked beats touch counts, replies need autoresponders stripped, and booked, held, accepted and qualified are four separate numbers.
Key takeaways
- Sort every number by who controls it before reading any of it. Reply rate, connect rate and meetings per contacted account measure the targeting decision, not the rep, and most unproductive reviews are those numbers discussed as though they were behavioural.
- Accounts genuinely worked is the honest activity measure and is rarely the one on the dashboard, because touches are easier to count. The gap between the two is where a struggling rep hides and where a good one is underestimated.
- Booked, held, accepted and qualified are four different numbers reported as one, and the distance between the first and the last is the real output of a sales development function.
- Published SDR benchmarks assume a multi-touch cadence and usually state no denominator. Under one message per campaign the denominator is people rather than touches, which makes the rates incomparable rather than merely different.
Reviewed and updated August 22, 2026
An SDR misses target for a second month and the review opens with activity. Dials are down, emails are flat, so the plan is more of both. Three weeks later the numbers look the same, because the accounts on the list had been worked twice already and no amount of additional contact was going to change that.
The problem is not that activity metrics are wrong. It is that a sales development scorecard usually mixes three kinds of number together, and only one of them is about the person being reviewed. Sorting them by who controls what turns the same dashboard into a diagnosis.
Three tiers of ownership
What the rep controls. How many accounts they worked, whether the research was done, what the message said, whether they followed up on commitments they personally made, how a conversation was run. These respond to coaching and they are the only ones that belong in a performance conversation.
What the list controls. Reply rate, connect rate, positive reply rate, meetings per contacted account. A rep working an exhausted or badly targeted segment produces worse numbers on all four while doing identical work. These measure the targeting decision, which is somebody else's job.
What the market and the infrastructure control. Bounce rate, spam placement, contact-data decay, seasonal buying patterns, whether a category is in a spending freeze. These arrive in the rep's figures and originate nowhere near them.
Almost every unproductive sales development review is a tier-two or tier-three number being discussed as though it were tier one.
- Accounts genuinely worked, not touched
- Research quality before the first message
- What the message actually says
- Commitments kept after a conversation
- How a conversation is run and closed out
- Reply rate and positive reply rate
- Connect rate and contact rate
- Meetings per contacted account
- Whether the premise is true of these accounts
- Bounce rate and spam placement
- Contact data decay in the segment
- Category-wide spending freezes
- Seasonality in the buying cycle
The metrics themselves, and what each is really for
Accounts worked. Distinct accounts that received genuine, researched contact in the period. This is the honest activity number and it is rarely the one on the dashboard, because touches are easier to count. The gap between touches and accounts is where a struggling rep is usually hiding, and where a good rep is usually being underestimated.
Emails sent and dials made. Compliance measures. They are worth watching for absence rather than for level: a rep whose activity collapses has a problem worth asking about, and a rep whose activity climbs after the number appears on a wall has responded rationally to being scored. Anything fully inside a rep's control degrades as a measure the moment it becomes the target.
Connect rate and contact rate. On calling, the share of dials that reach a human and the share that reach the intended person. These belong to the data and the segment more than to the rep, and they are the first thing to check when call outcomes move without any change in behaviour.
Reply rate, and positive reply rate separately. Total replies include out-of-office responses, refusals and bounced-back autoresponders, and a reply-rate figure that has not had those stripped out overstates by a large and variable margin. The two numbers move independently and a rise in one with a fall in the other is a specific and useful signal about the message.
Meetings booked, held, accepted and qualified. Four different numbers that get reported as one. Booked is a calendar event. Held is attendance. Accepted is a seller agreeing the meeting is worth their pipeline. Qualified is the meeting meeting a written standard. The distance between booked and qualified is the real measure of a sales development function, and the standard that makes the last one checkable is described in qualified appointment.
Pipeline created and its survival. The value of opportunities originating from sales development, and then the share of them still alive two stages later. The second half is what stops the first half being gamed, because pipeline created alone rewards volume of weakly qualified opportunities.
Bounce rate, complaint rate, unsubscribe rate. These are infrastructure metrics wearing a rep's name. They belong on a deliverability review rather than a performance one, and a rep penalised for them is being marked on somebody else's work.
The denominator most published SDR metrics assume
The standard sales development metric set was built for a multi-touch cadence. Reply rate means replies per sequence, meetings per touch assumes several touches per person, and the published figures circulating for all of these are computed on that basis.
RevenueFlow runs one message per campaign. No bumps, no thread replies, and a later approach is a separate campaign with a fresh premise rather than a follow-up to the first. That is our documented policy, and it changes the arithmetic in a way worth stating explicitly, because it makes our rates and the published ones incomparable rather than merely different.
The following figures are invented for the illustration and describe no campaign, ours or anyone's. Suppose a hundred people receive one message each and three reply. Suppose a hundred people receive four messages each in a sequence and four reply. Per person, the second is barely better. Per message sent, the first is four times the rate. Report those two setups against each other on either denominator alone and one of them looks decisively better, and neither reading is a fact about the sellers.
Illustrative figure
Illustrative figure
Direction is the point, not the size
So the practical rule is to fix the denominator per person and per account, state it beside every rate, and treat any external benchmark whose denominator is unstated as unusable. That last part removes most of the figures circulating for this function.
The two numbers usually missing entirely
Most sales development scorecards are complete on outputs and silent on two inputs that explain a large share of the variance in them.
Coverage. The share of the assigned list that has ever received real contact. Assignment gets recorded as coverage constantly, and they are different facts: a segment can be owned, researched, loaded into a tool and never messaged, and every report will show it as worked. A team whose coverage is at seventy percent has a different problem from a team whose coverage is complete and whose reply rate is poor, and the two are indistinguishable on any output metric.
Time from list to first contact. How long an account sits between being assigned and being contacted for the first time. This decays quietly, it is invisible on every activity count, and it is the input most sensitive to a rep being overloaded, because the natural response to too many accounts is to work the easy ones first and let the rest age. Where the contact is triggered by an inbound or event signal rather than by a list, the same delay is more expensive still, which is the argument in why the fastest rep to call wins the deal.
Both are cheap to compute and neither can be gamed by working harder, which is most of what recommends them.
What to review, and how often
Sales development produces more events than sales does, which makes weekly reading more defensible here than it is for closers. It does not make it safe on every metric.
Read weekly: accounts worked, whether messages went out at all, bounce and spam signals, and any conversation the rep flagged. All four are either large-sample or binary, and all four support a decision that week.
Read monthly or quarterly: reply rates by segment, meetings per contacted account, and anything about individual ranking. A single rep's meeting count in a single month is a small enough sample that ordinary variation moves it further than ability does, and the arithmetic behind that caution is worked through in sales rep performance metrics.
Never read as a rep's number at all: bounce rate, deliverability, and any rate whose denominator is a list they did not choose.
- Yes: Every rate on it states its denominator, per person or per message
- Yes: Replies are counted after out-of-office and autoresponders are stripped
- Yes: Booked, held, accepted and qualified are four separate columns
- Yes: Accounts worked appears alongside touch counts
- Depends: Deliverability numbers sit on an infrastructure review, not a rep's
- Depends: Individual rankings are read over quarters rather than weeks
- Depends: Someone checks whether a flat month is the list rather than the rep
Why no benchmark numbers appear here
Every published figure for this function that turned up while writing this page lives on a vendor or agency blog, states no denominator, and disagrees with the next one. A meetings-per-month figure computed on inbound-assisted sales development, on a multi-touch cadence, in one segment, is not a target for a team doing outbound cold with one message into another. Publishing it as a benchmark would make it one anyway.
The number worth having is your own, computed on your own denominator and read against your own previous quarter. Before it can carry any weight, the constraint has to be identified honestly, and the four candidates are the list, the message, the infrastructure and the coaching hours. That diagnosis is in outbound for SDR managers, and the part of it the manager personally owns rather than delegates is in what an SDR manager owns.
Where the honest answer is that the hours do not exist, the pricing shapes for buying the output instead are compared in outsourced SDR pricing.
The short version
Sort the scorecard by who controls each number before reading any of it. Accounts worked beats touches. Strip autoresponders before quoting a reply rate. Keep booked, held, accepted and qualified as four columns, because the distance between the first and the last is the function's real output. State the denominator beside every rate, and discard any external benchmark that does not.
If the constraint turns out to be the list rather than the team, we will build one campaign against a defined segment and you can read the replies yourself.
Frequently asked questions.
Frequently asked questions- What are the most important SDR metrics to track?
- Accounts genuinely worked, positive reply rate with autoresponders stripped out, and the four meeting numbers kept separate: booked, held, accepted by a seller, and qualified against a written standard. Add coverage, meaning the share of the assigned list ever contacted, because it explains more variance in the output metrics than anything else usually on the board.
- Should SDRs be measured on dials and emails sent?
- Watch them for absence rather than for level. A collapse in activity is worth asking about. A rise after the number appears on a wall is a rational response to being scored rather than an improvement, because anything fully inside a rep's control degrades as a measure the moment it becomes the target. Accounts worked is the more honest version.
- Why is reply rate misleading?
- Two reasons. Total replies include out-of-office responses, refusals and autoresponders, so an unstripped figure overstates by a large and variable margin. And the denominator is often unstated, so a rate computed per message in a multi-touch sequence gets compared against one computed per person. Track positive reply rate separately and state the denominator every time.
- What is a good number of meetings per month for an SDR?
- No publishable figure exists. Every benchmark located for this function sits on a vendor or agency blog, states no denominator, and disagrees with the next one. A figure computed on inbound-assisted sales development in one segment is not a target for an outbound team in another. Use your own number against your own previous quarter.
About the author.
B2B cold email experts helping companies generate qualified leads through done-for-you outreach campaigns.
RevenueFlow Team
Explore more.
Ready to scale your outreach?
We build GTM engines that book real meetings. See the receipts.
Related articles.
Sales Motivation Quotes: Seven Famous Lines, and Who Actually Said Them
Ford's most quoted line first appeared months after he died, with no source. Seven sales-floor staples, checked against the earliest printed record.
Outbound for SDR Managers: Finding the Constraint Before You Add Capacity
A short month is usually the list, the message or the infrastructure. Capacity is the least common cause and the only one more people actually fix.
Sales Rep Scorecard: What Belongs on It, and What the Rep Cannot Control
Most rep scorecards carry lines the rep cannot move. The controllability test, the six lines that earn a place, and the two ways the document stops meaning anything.
Cold Calling Sales Training: Diagnose Before You Buy It
Calling results come from the list, the offer and the execution. Training reaches the third, and the third is frequently not the one that broke.
SDR Role Definition: What Automation Left for the Person to Do
List building, enrichment, sequencing, data entry and first-draft personalisation have moved. What remains is judgment, which changes how you scope the seat.
Outbound SDR and Inbound SDR: One Title, Two Jobs
One rep decides which fifty companies are worth contacting today. The other answers overnight form fills. Same title, different constraints and fixes.