Questions d'entretien : Sales Manager

I start with productive capacity, then allocate quotas by territory potential rather than dividing the target equally.

• For six fully ramped reps at $800K each, gross capacity is $4.8M; I subtract expected vacancy and ramp loss before accepting a $4.5M team target.
• I weight assignments using reachable account value, historical conversion, and sales-cycle length so a mature enterprise patch is not treated like a new SMB patch.
• I keep a small coverage buffer, usually 10% to 15%, because 100% theoretical capacity leaves no room for attrition or uneven performance.

Sales capacity is the revenue a team can reasonably produce after accounting for headcount, productivity, and time in seat.

• A simple model is productive reps multiplied by annual quota multiplied by expected attainment, such as 8 x $1M x 85% = $6.8M.
• New hires count fractionally during ramp, so a rep at 25%, 50%, 75%, and 100% productivity across four quarters contributes $625K against a $1M quota.
• I model vacancies and planned leave separately because assuming every approved seat is productive overstates capacity.

A useful capacity model makes productivity losses visible instead of hiding them inside one attainment percentage.

• I include filled seats, hiring dates, ramp curves, voluntary attrition, expected attainment by tenure, and seasonality.
• I separate controllable assumptions, such as hiring date, from observed rates, such as 12% annual attrition, so owners are clear.
• I run base, upside, and downside cases because a two-month hiring delay can remove a full quarter of productive capacity.

I phase quota around when the team can actually close revenue, while preserving the annual commitment.

• If historical bookings split 20%, 25%, 25%, and 30% by quarter, a flat 25% allocation would distort both coaching and forecasting.
• A new rep starting in Q1 might carry 0%, 50%, 75%, and 100% of a full quarterly quota as product knowledge and pipeline mature.
• I avoid making Q4 the plug for every earlier shortfall because that creates an impossible finish and encourages low-quality deals.

Pipeline coverage is qualified pipeline divided by remaining quota, but the right multiple comes from conversion rather than a universal rule.

• If the team wins 25% of qualified pipeline, the mathematical baseline is 4x coverage.
• I add a buffer for slippage and data quality, then compare the resulting 4.5x or 5x target with available selling capacity.
• I calculate it against the same time window and value basis, such as Q3 ARR pipeline divided by Q3 remaining ARR quota.

Coverage should reflect each segment's win rate, cycle, and slippage profile.

• An SMB motion winning 30% of opportunities may need about 3.3x qualified coverage, while enterprise at 15% needs about 6.7x.
• Enterprise pipeline must be created earlier because a 180-day cycle cannot repair a current-quarter gap, even when nominal coverage looks high.
• I monitor new business, expansion, and partner pipeline separately because their conversion rates are not interchangeable.

Unweighted pipeline shows total opportunity value, while weighted pipeline applies a probability to each opportunity.

• A $200K deal at 40% contributes $80K to weighted pipeline but the full $200K to unweighted pipeline.
• Unweighted value is better for capacity and coverage visibility; weighted value is more useful as one forecasting input.
• Neither replaces deal inspection because a stale late-stage deal can inflate both measures.

Stage weighting is easy to calculate, but it treats unlike deals as if stage alone determined probability.

• Two proposal-stage deals may differ sharply if one has an identified economic buyer and the other has only a friendly user.
• Fixed weights also lag changes in segment, source, rep tenure, and competitor, so 60% may be wrong for the current cohort.
• I calibrate weights from recent closed cohorts and pair the result with qualification evidence and days in stage.

I compare each opportunity's time in stage and total age with the normal range for its segment.

• If enterprise evaluation usually lasts 30 days, a deal sitting there for 75 days needs new evidence or a lower forecast category.
• I use cohort percentiles rather than one company-wide cutoff because a 45-day SMB deal and a 45-day enterprise deal mean different things.
• Aging should trigger inspection, not automatic closure, since procurement or legal can create legitimate delays.

Forecast categories need observable buyer evidence and a consistent confidence meaning.

• Pipeline means qualified but not expected this period; best case means plausible with named gaps; commit means the rep and manager expect it to close in period.
• A commit deal should have a confirmed problem, decision process, commercial path, and dated next step, not just a late CRM stage.
• I publish exit criteria and sample deals so every rep uses the labels the same way.

Stages describe where a buyer is in the sales process, while forecast categories express confidence about timing and outcome.

• A contract-stage opportunity can remain best case if security approval has no date, despite being late in the process.
• A discovery-stage expansion may be commit when the order is routine and the customer has confirmed budget and signature timing.
• I keep the fields separate so reps do not advance stages merely to improve the forecast.

I triangulate several methods because each fails in a different way.

• Stage-weighted forecasting multiplies value by historical stage conversion, while cohort forecasting applies conversion and timing from similar segment and source cohorts.
• Run-rate forecasting is useful for high-volume SMB motions, such as bookings to date divided by elapsed selling days multiplied by total selling days.
• I compare those outputs with manager judgment based on deal evidence rather than averaging them blindly.

I build the forecast from inspected opportunities, then reconcile it with historical conversion and capacity.

• Each rep submits commit and best case with amount, close date, next step, and the main unresolved risk.
• I challenge category changes using buyer evidence, then sum the accepted deals and add a modeled amount for repeatable high-volume business if appropriate.
• If the result is far above the cohort model, I identify the exact deals causing the gap rather than forcing the numbers to match.

I measure both the size and direction of forecast error at consistent snapshots.

• Forecast error can be calculated as absolute actual minus forecast divided by actual; $900K forecast versus $1M actual is 10% error.
• I also track bias, because repeated under-forecasting and over-forecasting can have the same absolute error but require different coaching.
• I compare week-4, week-8, and final commit accuracy by rep and segment so late accuracy does not hide poor early visibility.

Slippage is opportunity value pushed from its expected close period into a later one.

• I calculate slipped value divided by opening-period pipeline and split it by stage, segment, and rep.
• Repeated movement without a buyer-confirmed event signals wishful close dates, while one legal delay may be legitimate.
• I preserve close-date history in the CRM because overwriting the date destroys the evidence needed to diagnose the pattern.

I decompose the funnel before blaming lead quality or rep effort.

• I compare stage-to-stage conversion by source, segment, cohort month, and rep tenure to find where the break begins.
• If meeting-to-opportunity holds but proposal-to-win falls, I inspect qualification, competition, pricing, and decision process rather than top-of-funnel volume.
• I use mature cohorts so newly created opportunities are not mislabeled as losses simply because they have not had time to close.

Win rate is won opportunities divided by closed opportunities for a defined cohort, with consistent inclusion rules.

• I state whether no-decisions count as losses because excluding them can make a weak process look healthy.
• I segment by deal size, source, product, and competitor, since a blended 24% rate can hide 35% SMB and 10% enterprise performance.
• Revenue win rate complements logo win rate when a few large deals materially change the business outcome.

Sales velocity estimates revenue produced per unit of time from opportunities, deal size, win rate, and cycle length.

• The common formula is opportunities x average deal value x win rate divided by average sales-cycle days.
• With 40 opportunities, $25K average value, 20% win rate, and a 50-day cycle, velocity is $4K per day.
• It is a diagnostic model, not booked revenue, because averages can hide segment mix and extreme deals.

I calculate every velocity input by segment and compare it with a prior mature cohort.

• Fewer opportunities points to pipeline creation, lower deal size to account mix or discounting, lower win rate to qualification or execution, and longer cycle to process friction.
• I estimate impact before acting: improving win rate from 20% to 22% raises modeled velocity 10% if other inputs hold.
• I avoid optimizing one lever in isolation because pushing more weak opportunities can reduce win rate and consume rep capacity.

Cohorts keep unlike opportunity populations and maturity windows from being mixed together.

• I group by creation month or quarter, then compare outcomes only after enough time has passed for that segment's normal cycle.
• This prevents a surge of fresh pipeline from depressing apparent win rate or shortening cycle time among the few deals already closed.
• I also split by source and segment when mix changes, because a partner-led enterprise cohort should not benchmark inbound SMB.