How should a 2027 sales org design role cards for AE BDR CSM and RevOps?

The Tuesday morning that exposes a missing role card

Picture a Series B software company with twenty-two people carrying a number. Eight account executives, six business development reps, five customer success managers, three people in revenue operations. Annual recurring revenue somewhere in the mid-teens of millions. The company just closed a funding round on the strength of a hiring plan, so the headcount is about to grow by half again inside two quarters. On paper the org chart is clean. In practice, on a Tuesday morning in the middle of the quarter, four separate conversations are happening that all trace back to the same missing artifact.

In the first conversation, a BDR who has been on the floor for nine weeks is asking a manager whether she is allowed to disqualify an account that a marketing scoring model flagged as hot. She thinks the fit is wrong — wrong employee count, wrong buying trigger, no budget authority in the contact she can reach. Her manager says "use your judgment," which sounds empowering and is actually an abdication, because nobody has ever told her where her judgment ends and someone else's begins. She works the account for another eleven days, books a meeting out of politeness, and the AE burns forty minutes discovering what she already suspected in week one.

In the second, an account executive is sitting on a deal that needs a nineteen percent discount to close before the quarter ends. He does not know whether that requires his director, the VP, or finance. He asks in a Slack channel. Two people answer differently. He waits two days for a decision that a written threshold would have made in four seconds.

In the third, a CSM inherits a book of seventy-one accounts from a departing teammate and has no defined first-thirty-days motion. She starts with the loudest customers, which is the natural human instinct and exactly the wrong prioritization, because the loudest customers are rarely the ones silently drifting toward non-renewal.

In the fourth, the newest revenue operations hire has been asked to "clean up the pipeline data" — a request with no definition of done, no quality target, and no stated authority to change a single field-level validation rule without three approvals.

None of these four people is underperforming. Each of them is operating without a written statement of what they own, what they may decide alone, and what a good ninety days looks like. That statement is the role card. It is not a job description. A job description is written to attract a candidate and is functionally dead the moment the offer is signed. A role card is written to run the job and is referenced weekly for as long as the person holds the seat. Once a sales org crosses roughly fifteen to twenty quota-carrying people, the informal transmission of these norms — the manager who explains discounting authority over lunch — stops scaling, and the gaps show up as escalation latency, inconsistent qualification, and first-year attrition that leaders misdiagnose as a hiring problem.

How the eight-block card actually works

The design principle is that the skeleton never varies and the contents always do. Every card in the company has the same eight blocks in the same order, so a manager can scan any role's card in thirty seconds and a cross-functional partner can find the escalation line without reading prose. Here is what each block does and how to write it.

Purpose is one sentence, present tense, naming the outcome the seat exists to produce. "Close net-new revenue in a defined territory." "Generate qualified pipeline through outbound prospecting." "Protect and grow revenue in a defined book of business." "Build and operate the systems that let go-to-market teams run at scale." If the purpose sentence needs an "and" plus a semicolon, the seat is probably two seats.

Primary KPI is exactly one number. This is the hardest discipline in the whole exercise, because every leader wants three. One number means that when the person has ninety minutes and two competing tasks, the tiebreak is unambiguous. For an AE it is net-new bookings against quota. For a BDR it is qualified opportunities accepted by sales per month — accepted, not booked, so the BDR and AE share a definition and the BDR cannot win by setting meetings nobody wants. For a CSM it is net revenue retention across the book. For a RevOps individual contributor it is harder and worth thinking about carefully; the honest version is usually a service-level number the team commits to, such as the share of requested analyses delivered within a stated turnaround, or a data-completeness percentage on the fields that forecasting depends on.

Secondary KPIs are two or three metrics that constrain how the primary number may be earned. They exist to prevent the primary from being gamed. An AE whose only metric is bookings will discount to the floor, so pair bookings with average discount depth and win rate. A BDR measured only on opportunity count will spray, so pair it with the meeting-to-accepted-opportunity conversion rate. A CSM measured only on net retention can hide churn behind one large expansion, so pair net retention with gross retention. This pairing logic is the most transferable idea on the card, and it is the same logic that makes a good compensation plan: never pay for a number that can be reached by a behavior you would fire someone for.

Daily activities is five to eight concrete verbs with counts attached. Not "prospect consistently" but "complete a defined number of personalized first-touches before noon." Not "maintain pipeline hygiene" but "update next steps and close dates on every open opportunity before the Friday forecast call." The counts should reflect what your top performers actually do, measured, not what a leader believes they should do. Pull the numbers out of your own activity data rather than importing them from a benchmark report; the benchmark is for sanity-checking, not for setting.

Decision rights is the block that eliminates the most wasted hours, and the one most orgs skip. Write it as a list of things this person may do without asking anyone: discount up to a stated threshold, commit a stated number of solutions-engineering hours per opportunity, choose accounts within an assigned tier, schedule an executive business review, change a report definition inside a governed namespace. Every item is a decision the person will face in a normal week.

Escalation triggers is the mirror image: the specific conditions that require pulling someone else in, with the named role and a time bound. Discount above the threshold goes to the director. Any non-standard legal or security term goes to the director plus legal. A deal past a stated age with no next step goes to a structured review. An account health score below a defined floor goes to the director and the executive sponsor within a week. A production system outage past an hour goes to the head of go-to-market. The time bound matters as much as the trigger; "escalate at-risk accounts" without a clock produces escalations that arrive after the renewal is already lost.

Milestones at thirty, sixty, ninety, and one hundred eighty days convert the abstract idea of ramp into a checkable sequence. The rule for writing them is that each milestone must be verifiable by someone other than the person being measured. "Understands the product" is not a milestone; "passes the certification and delivers a recorded demo scored by a manager" is.

Tools names three to seven systems the person must be able to operate, with the specific in-house workflow attached rather than just the vendor name. "The CRM" is useless; "the CRM, specifically opportunity creation, the stage-exit criteria, and the forecast category field" is a training plan.

The second design principle is that the card must live where the work happens. A role card stored in a shared drive folder is a document; a role card referenced in the offer letter, pinned to the CRM home view, embedded as the agenda skeleton of the weekly one-on-one, and quoted in the compensation plan is an operating system. The test is simple: if you deleted every copy of the card tomorrow, how many meetings would notice? If the answer is none, you wrote a job description with extra steps.

What goes on each of the four cards

The four seats differ enough that generic guidance fails, so here is the shape of each, with the reasoning behind the choices rather than a set of numbers to copy blindly.

The account executive card turns on segment. A mid-market AE running a sixty-to-ninety-day cycle and an enterprise AE running six to twelve months should not share a card, even though both say "AE" on the org chart. The mid-market card is activity-dense: several discovery calls and demos a week, an explicit prospecting obligation whenever pipeline coverage falls below the stated multiple of quota, daily hygiene. The enterprise card is account-plan-dense: a smaller number of active pursuits, a required multi-threading count per opportunity, a written account plan reviewed quarterly, and a stakeholder map that has to name economic buyer, champion, and likely blocker. Pipeline coverage targets differ too — a shorter cycle with a higher win rate needs less coverage than a long cycle with a lower one, and the right multiple falls out of your own historical conversion math rather than a rule of thumb. Decision rights on the AE card should always include a discount threshold and a solutions-engineering budget, because those are the two resources an AE spends daily that belong to someone else.

The business development card has the sharpest handoff problem, so its escalation block does the heaviest lifting. Two triggers matter most: what happens when a prospect asks a pricing or security question the BDR is not authorized to answer, and what happens when the rep goes an extended stretch without producing an accepted opportunity. The first needs a stated handoff window measured in hours. The second is a coaching trigger, not a discipline trigger, and the card should say so in plain language — otherwise the escalation block reads as a threat and the rep hides the problem. The BDR card is also where a definitional dependency lives: the accepted-opportunity criteria must be written identically on the BDR card and the AE card. If the two cards disagree by even one clause, you have manufactured a recurring argument. Segment matters here too. An inbound-response rep working speed-to-lead has a fundamentally different daily activity block than an outbound rep working a named-account list, and pretending otherwise produces a card neither one can follow.

The customer success card is the one most often written as a feelings document. Resist that. The primary metric is retention of revenue, and the card should name the book construction rule that makes it achievable — how many accounts, at what revenue band, with what tiering of touch. A CSM carrying eighty accounts cannot run the same motion as one carrying twelve, so a single card covering both is a card covering neither. The daily activity block should distinguish proactive touches from reactive ones and require a floor on the proactive count, because reactive work will otherwise consume the entire week. The decision rights block should explicitly grant the authority to escalate an at-risk account without a manager's pre-approval; making a CSM ask permission to raise an alarm is how renewals get lost quietly. Milestones should include inheriting the book, contacting every named champion, producing a written risk assessment, and delivering a first business review.

The revenue operations card is the one most orgs get wrong, usually by writing a ticket queue instead of a role. RevOps is not an internal help desk, and if the card reads like one, that is what the seat becomes. Split the daily activity block explicitly into run and build: a bounded share of time on requests and system maintenance, a protected share on the roadmap of improvements. Without that split, the queue eats everything and the systems never get better. Decision rights are where a RevOps card earns its keep — this person needs written authority to change configuration inside a governed area, to define and version a metric, and to decline a request that violates the data model. That last one is the single most valuable line on any RevOps card. The escalation block should cover outages, compensation calculation errors, and vendor commitments above a threshold, each routed to a named finance or go-to-market leader. If your RevOps function has more than a couple of people, split the cards further by discipline — systems administration, analytics and reporting, and enablement or process design all belong on separate cards, because the tools block alone diverges completely between them.

Two cross-cutting elements deserve a place on every card. The first is a short accountability line naming what this role owns, supports, and is merely informed about at each stage from prospecting through onboarding and expansion. Writing that line for all four seats at once is a useful forcing function — it surfaces every gap where two roles both think the other owns something, and every overlap where both are doing the same work. The second is a small set of behavioral anchors: three to five one-sentence statements of how the work is done, each tied to an activity or decision right already on the card. "Documents discovery in the CRM the same day" is an anchor. "Is a team player" is not.

Reading the numbers without importing someone else's

Every benchmark report you will encounter has the same failure mode when applied to role cards: it reports a median across a population whose segment mix, deal size, motion, and tenure distribution look nothing like yours. Use benchmarks as guardrails, not as targets.

The right way to set the numbers is to derive them from your own data first, then check them against the outside world. Start with quota, which is a capacity calculation rather than an aspiration: take your median closed-won deal size over the last four quarters, take the win rate from the stage where an opportunity becomes real, take the average cycle length, and compute how many opportunities a fully ramped rep can carry simultaneously given the touch time each one requires. That produces a defensible bookings number and, working backward through the conversion rates, the pipeline coverage multiple the card should state. If your win rate from qualified opportunity to closed-won is one in four, you need meaningfully more coverage than an org converting one in three, and no report can tell you which one you are.

Do the same for the BDR card. Count the qualified opportunities your AEs actually need per month to hit the coverage target, divide across the BDR headcount, and then check whether the implied meeting count is reachable at your historical touch-to-meeting rate. Very often it is not, and the answer is either more reps, better data, or a lower quota — but the card should never state a number the math says is impossible, because that turns the card from a contract into a fiction and everyone stops believing the rest of it.

For CSM book construction, work from touch capacity. Decide how many proactive touches a business review cycle requires per account per quarter, multiply by book size, add reactive load estimated from your own ticket and escalation history, and see whether the total fits a working week. If it does not, either the book is too big or the touch model is too rich, and the honest move is to tier the book explicitly on the card — a small number of high-touch accounts, a larger pooled tier, and a digital-only tail.

For RevOps, the number that matters most is the run/build split, and the way to set it is to measure it. Track a few weeks of actual time against inbound requests versus roadmap work. Most teams discover the split is far more lopsided toward run than anyone believed, and the card is the instrument for correcting it — write the target split, then enforce it with a request intake process rather than heroics.

Where external benchmarks genuinely help is in sanity-checking direction. If your computed AE quota is triple the median for your segment and deal size, something in the math is wrong or the hiring plan is fantasy. If your ramp milestones have a rep at full productivity in sixty days while comparable orgs take two or three quarters, your milestones will read as insulting to anyone who has done the job. Look at published sales-development and sales-effectiveness benchmarks for that directional check, and look at your own attrition and ramp data for the actual targets.

One more measurement note. Instrument the card itself. Once a quarter, ask every person who holds one to answer three questions in a survey: can you state your primary metric without looking it up, do you know the largest discount or change you can approve alone, and do you know the trigger that requires you to escalate. The share answering yes to all three is the only real measure of whether the cards are working. A low score is not a sign the cards are wrong; it is a sign they are not living anywhere people look.

Trade-offs worth arguing about before you publish

Uniform skeleton versus role-specific format. Uniformity is what makes handoffs legible and what lets a manager who moves from the BDR team to the CSM team read a card cold. The cost is that some blocks feel forced — a RevOps analyst's "daily activities" block genuinely is lumpier than a BDR's. Keep the skeleton anyway and let the contents flex. The moment you allow one team to invent its own format, comparison across roles dies and so does the handoff mapping.

Detail versus adoption. A three-page card is more complete and less read. The one-page constraint is not aesthetic; it is the mechanism that forces prioritization. If the discount threshold does not fit on the page, something less important is taking its place. Push supporting detail into linked playbooks and keep the card itself scannable in under a minute.

Annual refresh versus quarterly. There is a real tension here and both positions have merit. Annual refresh at planning keeps the card stable, which matters enormously for anything tied to compensation — changing a KPI mid-quarter breaks accrual logic and reads to sellers as moving the goalposts. Quarterly refresh keeps the card current in a fast-changing stack, and stale cards get ignored. The resolution most orgs land on is a split cadence: the compensation-bearing blocks — primary KPI, secondary KPIs, decision-right thresholds — change only at the annual planning cycle. The operational blocks — daily activities, tools, milestones — can be revised quarterly. Version-stamp the card so everyone can see which revision they are reading.

One card per role versus one per role-segment. Segmentation is almost always correct and almost always resisted, because it multiplies the maintenance burden. The compromise that works is a shared core with a segment appendix: identical purpose, identical block structure, identical decision-right categories, with segment-specific numbers in a clearly marked band. That keeps the count of genuinely distinct documents manageable while acknowledging that enterprise and SMB are different jobs.

Who owns the card. RevOps owning the template and the process, with the functional leader owning the content, is the arrangement that survives contact with reality. RevOps alone produces cards that are internally consistent and operationally naive. The functional leader alone produces cards that contradict each other at the handoffs. Human resources should touch only compliance language. Involving one top performer per role in the drafting is worth the hour — they will tell you which of your proposed daily activities nobody actually does.

The failure modes that make role cards useless

Writing them as job descriptions. The tell is aspirational verbs with no object: drives, owns, partners, champions. If a line does not name a number, a threshold, a system, or a named person, cut it. A useful heuristic: read each line and ask what a person would do differently tomorrow morning because of it. If the answer is nothing, it is decoration.

Skipping decision rights. This is the most common omission and the most expensive one, because it is invisible. Nobody files a complaint titled "I waited two days for an answer I should have been allowed to give myself." The cost shows up as cycle time and as managers whose calendars are full of approvals. Every card needs the section, and the thresholds need to be numbers, not adjectives. "Reasonable discounts" is not a decision right.

Escalation triggers with no clock and no name. "Escalate when the deal is at risk" delegates the definition of risk to the person least equipped to judge it. Write the observable condition and the deadline: this state, to this role, within this many hours or days.

Milestones nobody can verify. If a manager cannot look at a system and determine whether the ninety-day milestone was hit, it is not a milestone. Tie each one to an artifact — a recorded call, a created opportunity, a delivered business review, a shipped dashboard.

Changing cards mid-quarter. Even a well-intended clarification lands as a goalpost move when compensation depends on it. Batch changes into the published windows and communicate them the way you would a plan change, with a written summary of what moved and why.

Letting the card contradict the compensation plan. This is the credibility killer. If the card says the primary metric is net revenue retention but the compensation plan pays on new logo expansion only, the card loses every argument. Do the reconciliation explicitly: lay the four cards and the four plans side by side, and confirm that each primary KPI appears as a material component of the matching plan.

Never mapping the handoffs. Individual cards written in isolation produce seams. Build the accountability matrix as a single exercise across all four roles, and specifically nail the three seams that generate the most friction: BDR to AE at opportunity acceptance, AE to CSM at closed-won, and everyone to RevOps for data and reporting. For each seam, write what transfers, in what format, within what window, and what the receiving role may reject it for. The right to reject a bad handoff is what makes the handoff standard real.

Publishing and walking away. A card that is not referenced in the one-on-one is dead within a quarter. The lowest-cost enforcement mechanism is to make the card the agenda: milestones for anyone in ramp, primary and secondary metrics for anyone past it, and a standing question about whether any escalation trigger fired this week. That takes about six minutes and it is the entire maintenance cost of the system.

Ignoring the adjacent seats. Once the four core cards exist, the same skeleton extends naturally to sales engineering, partner managers, renewals specialists, and first-line managers — and the manager card is worth writing early, because the manager's decision rights are the ceiling on everyone else's escalations. If the AE card says discounts above a threshold go to the director, the director's card had better state what the director may approve alone.

Related questions

Should first-line managers get role cards too?

Yes, and ideally before the individual contributor cards are finalized. Manager decision rights define the ceiling on every escalation path below them. If an AE escalates a discount to a director who also lacks authority, you have added a hop without adding a decision.

How do role cards interact with career ladders?

They are complementary but distinct. A role card describes the current seat's operating reality; a ladder describes progression across seats. The cleanest link is to reference the next level's card in the milestone block, so the path forward is a document rather than a conversation.

What if a rep's actual work does not match the card?

Investigate before you correct. Frequently the card is wrong — it describes what leadership imagines rather than what produces results. Compare the top performers' actual activity data to the card and update the card if reality is outperforming the document.

Do smaller teams need this?

Under roughly ten quota carriers, informal transmission usually works and cards feel like bureaucracy. Write them anyway, briefly, before the growth spurt. Retrofitting cards onto forty people who each learned a different informal norm is far harder than starting with eight.

Can the same approach work outside of sales?

The eight-block structure transfers well to any operationally-dense function — support, implementation, marketing operations. Decision rights and escalation triggers are the portable parts; the metric blocks need rethinking wherever individual output is harder to isolate.

FAQ

What is the difference between a role card and a job description?

A job description is a recruiting artifact optimized to attract applicants, and it usually stops being read the day the offer is accepted. A role card is an operating artifact optimized to run the job: it names one primary metric, the decisions the person may make alone, the conditions requiring escalation, and verifiable ramp milestones. The job description gets the person in the door; the role card tells them what to do on Tuesday.

How long should a role card be?

One page, hard constraint. The page limit is the forcing function that makes leaders decide what actually matters. Anything that does not fit belongs in a linked playbook rather than on the card. If a reader cannot scan the whole thing in under a minute and come away knowing the primary metric and the escalation path, it is too long.

Should AE cards differ by segment?

Yes. An enterprise seat with a multi-quarter cycle and a mid-market seat with a sub-quarter cycle are different jobs that happen to share a title. Keep the eight-block skeleton identical and vary the numbers and activities in a clearly marked segment band. That preserves comparability without pretending the roles are the same.

Who should own writing and maintaining them?

Revenue operations owns the template, the version control, and the refresh process; the functional leader owns the substance for their team. Bring in one top performer per role during drafting to sanity-check the daily activity block. Human resources reviews only for compliance and legal language — content ownership sitting in HR is the fastest route to cards nobody uses.

How do role cards connect to compensation plans?

The card's primary KPI should appear as a material component of the matching compensation plan, and the secondary KPIs should map to whatever accelerators, gates, or modifiers exist. Do the reconciliation as an explicit side-by-side exercise. When a card and a plan disagree, people follow the plan and stop trusting the card.

Can role cards be used in performance reviews?

They should be the backbone of the review, because they were written in advance and they are specific. Milestones give the ramp conversation an objective structure, decision rights and escalation triggers give the judgment conversation concrete reference points, and behavioral anchors give the coaching conversation language that is not purely numeric. Using a document written before the period being reviewed also removes most of the recency bias from the conversation.

Sources

• https://hbr.org/2017/07/what-your-sales-organization-can-learn-from-the-way-startups-sell
• https://www.gartner.com/en/sales/topics/sales-operations
• https://www.salesforce.com/resources/research-reports/state-of-sales/
• https://www.bridgegroupinc.com/blog
• https://blog.hubspot.com/sales/sales-compensation
• https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights
• https://www.gainsight.com/blog/
• https://openviewpartners.com/blog/

Related on PULSE

• [How do you create dynamic battle cards that update from competitor news feeds?](/knowledge/q9801)
• [How does the 2027 buying committee's preference for self-service research affect the role of the BDR?](/knowledge/q16359)
• [My company eliminated the BDR role — what should I do?](/knowledge/q1483)
• [When does adding a sales operations BDR (admin assistant for reps) actually free up real selling time?](/knowledge/q218)
• [What is Artisan (Ava) and why is it a hot RevOps AI BDR for 2027?](/knowledge/q12217)