One-time audit · Jan 1 – Jul 6, 2026 · ActiveCampaign

YTD Sales Discipline: five pipelines, ten agents

Every deal, activity log, and task pulled from ActiveCampaign for NewLeads, TarikGidamyPipeline, HarbourTEN10, TheBedford and 8Temple — 3,546 deals, then de-noised: automation touches and the 4,133 lead-intake notes the integration writes under the shared admin login are excluded, so every number below reflects verifiable human work.

3,546deals in scope YTD
4.9 hmedian speed to lead, in 9am–9pm working hours (14.6 h raw)
22%of new leads get a first touch within 1 working hour
2.5avg human attempts per lead
75%of 2,068 open deals untouched 30+ days
3deals marked Won YTD

Speed to lead measured on the 3,039 leads created YTD (TheBedford excluded — its leads predate 2026). Working-hour figures count only time elapsed between 9am and 9pm Toronto time, so overnight leads aren't penalized.

What the data says

Ordered by how much they should influence your agent and system decisions.

Systemic

Speed to lead is the org's biggest leak — median 4.9 working hours, only 22% of leads touched inside one

Even counting only the 9am–9pm window (no penalty for overnight leads), the first human touch on a new lead takes a median of 4.9 working hours (14.6 h on the raw clock). Arissa proves the ceiling: 13 minutes, 83% inside one working hour — on the same lead flow. Everyone else sits between 1.1 (Penny) and 15 (Nikhil) working hours. This is a process gap, not a lead-quality gap.

Model

Arissa Moreau is the discipline benchmark — and her system is the task queue

13-minute median first touch (working-hours clock) on NewLeads, 8.1 attempts per lead across her 373 deals, only 28% of her open NewLeads stale, and 1,880 tasks worked to an 87% completion rate with 2 overdue. She is the only agent whose CRM behavior is task-driven end to end (dark teal in the style column). Her Bedford follow-up load (8.3 attempts/lead on 3-year-old leads) is the hardest work in the dataset. Whatever cadence she runs is your template for everyone else.

Agent

Nikhil Oberoi: 37% of his 309 YTD leads were never touched by a human; 8Temple is 50% untouched

Across his NewLeads and 8Temple lead flow he had a 15-working-hour median first touch (more than a full selling day; 31 h raw) and 113 leads with zero human activity, ever (24% of his NewLeads, 50% of 8Temple). On 8Temple (150 leads, his alone) half were never worked and 121 open tasks are past due. His attempts average is 0.6–1.4 per lead against Arissa's 8.1. This is the clearest underworking signal in the data.

Pipeline

HarbourTEN10 is on life support: 95% of its 462 open deals are 30+ days cold

Elena's 285 open deals there are 100% stale with 377 overdue open tasks; Salim's side is 88% stale but his task hygiene is far better (1,767 of 1,948 done, 21 overdue). Decide whether this project is being actively sold or formally parked — right now it's parked in all but name.

Agent

Tarik works his 1,084 leads via stage moves and notes — with no task system and an 8.2-working-hour first touch

Only 7% of his leads get touched inside a working hour. He completed 29 tasks all year (Arissa: 1,115) but moved stages ~1,900 times — he manages the board, not a follow-up queue. He also does much of his work logged in as the shared admin account (2,097 of its 2,770 work actions are in his pipeline), which makes his activity only partially auditable.

System

The shared admin login pollutes attribution — 6,900+ actions can't be tied to a person

4,133 of its notes are the lead-intake integration (fine, excluded here), but 2,770 are real follow-up work done by someone logged in as admin. If you want accountability metrics to mean anything going forward, every human needs their own seat, and the integration needs its own API user.

System

Three deals marked Won all year, against 1,475 Lost — closes aren't being recorded

Either nothing closed YTD across 3,546 deals (unlikely) or agents mark losses but abandon deals when they win. Every conversion metric you might want later is unusable until wins are logged.

Style

You have four different CRMs in one: task users, note takers, stage movers, and email-only agents

Arissa runs on tasks; Matthew logs everything as notes (399 notes, zero task completions — but 0% of his leads went untouched); Tarik and Rex move stages; Elena, Ashley and Penny lean on synced email. None of these styles are wrong per se, but they are unauditable against each other — pick one canonical "attempt" (task or note) and make it policy.

Agent scorecard — all pipelines combined

Attempts = notes + completed tasks + synced emails by that agent's own login. "Attempts on lost" shows how hard leads are worked before being marked lost.

AgentLeadsSTL median (work hrs)<1 wh<4 whNever touched Attempts avgon lostStale 30 d+Tasks doneOverdue openWork style
Tarik Gidamy 1084 8.2 h19.6 h raw 7% 26% 2% 1.5 1.3 65% of 742 39% 127
Salim Walji 566 3.6 h12.8 h raw 29% 53% 8% 2.3 2.0 80% of 204 90% 29
Elena Novikova 500 3.6 h11.5 h raw 27% 52% 6% 2.0 1.5 86% of 398 57% 475
Arissa Moreau 373 13 min19 min raw 83% 95% 1% 8.1 3.7 57% of 242 87% 2
Nikhil Oberoi 309 15.0 h31.3 h raw 6% 25% 37% 1.0 1.1 94% of 193 63% 188
Matthew Dorosh 250 3.2 h12.3 h raw 17% 55% 0% 2.7 2.0 91% of 67 95% 17
Penny Ginsberg 211 1.1 h1.7 h raw 48% 66% 9% 3.0 2.2 86% of 133 72% 92
Ashley Tsitopoulos 121 3.6 h5.2 h raw 34% 51% 15% 1.8 1.7 79% of 19 94% 16
Rex Romero 109 9.2 h21.2 h raw 11% 28% 1% 0.9 0.5 96% of 55 87% 33
Joanne Lee 10 26 min59 min raw 67% 89% 10% 1.0 1.5 38% of 8 74% 5
Task completionsNotes Synced emailsStage moves

By pipeline

PipelineDealsSTL median (work hrs)<1 wh<4 whNever touched Attempts avgon lostStale 30 d+Tasks doneOverdue openWork style
NewLeads 1184 2.6 h9.2 h raw 35% 58% 8% 2.8 1.8 69% of 523 86% 326
TarikGidamyPipeline 1084 8.2 h19.6 h raw 7% 26% 2% 1.5 1.3 65% of 742 39% 127
HarbourTEN10 875 3.7 h12.4 h raw 27% 52% 6% 2.3 2.0 95% of 462 79% 403
TheBedford 253 n/a 5% 7.3 5.0 73% of 211 81% 13
8Temple(S.P.) 150 16.2 h38.5 h raw 3% 16% 50% 0.6 0.6 92% of 130 44% 121

TheBedford has no 2026 lead flow — its numbers cover YTD follow-up work on 253 pre-existing deals, so speed to lead doesn't apply.

By pipeline × agent

Agent · pipelineLeadsSTL median (work hrs)<1 wh<4 whNever touched Attempts avgon lostStale 30 d+Tasks doneOverdue openWork style
Nikhil Oberoi8Temple(S.P.) 150 16.2 h38.5 h raw 3% 16% 50% 0.6 0.6 92% of 130 44% 121
Salim WaljiHarbourTEN10 507 3.9 h13.6 h raw 27% 51% 8% 2.3 2.0 88% of 173 91% 21
Elena NovikovaHarbourTEN10 363 3.2 h10.9 h raw 27% 54% 4% 2.4 1.7 100% of 285 51% 377
Matthew DoroshNewLeads 250 3.2 h12.3 h raw 17% 55% 0% 2.7 2.0 91% of 67 95% 17
Arissa MoreauNewLeads 175 13 min19 min raw 83% 95% 1% 7.8 3.5 28% of 69 93% 2
Nikhil OberoiNewLeads 159 11.6 h23.6 h raw 8% 31% 24% 1.4 1.2 97% of 63 77% 67
Penny GinsbergNewLeads 158 1.1 h1.7 h raw 48% 66% 5% 2.7 1.3 82% of 95 75% 80
Elena NovikovaNewLeads 137 5.8 h15.3 h raw 26% 46% 12% 1.2 0.8 50% of 113 70% 98
Ashley TsitopoulosNewLeads 121 3.6 h5.2 h raw 34% 51% 15% 1.8 1.7 79% of 19 94% 16
Rex RomeroNewLeads 109 9.2 h21.2 h raw 11% 28% 1% 0.9 0.5 96% of 55 87% 33
Salim WaljiNewLeads 59 59 min1.7 h raw 50% 80% 15% 2.1 1.9 39% of 31 81% 8
Joanne LeeNewLeads 10 26 min59 min raw 67% 89% 10% 1.0 1.5 38% of 8 74% 5
Tarik GidamyTarikGidamyPipeline 1084 8.2 h19.6 h raw 7% 26% 2% 1.5 1.3 65% of 742 39% 127
Arissa MoreauTheBedford 198 n/a 1% 8.3 4.6 68% of 173 82% 0
Penny GinsbergTheBedford 53 n/a 19% 3.7 6.4 95% of 38 42% 12

Explore: build your own cut

Every metric recomputes live from the 3,546 underlying deals. Filter, then group the result however you need. Small groups (under ~20 leads) are directional at best.

GroupLeadsSTL median (work hrs)<1 wh<4 whNever touched Attempts avgmedon lost0-attempt6+Stale 30 d+Tasks doneOverdue openWork style
Task completionsNotes Synced emailsStage moves

Method & caveats

Scope. Deals created Jan 1 – Jul 6, 2026 in NewLeads, TarikGidamyPipeline, HarbourTEN10 and 8Temple(S.P.); for TheBedford, the 253 pre-2026 deals with YTD activity. Full activity logs and all 9,343 linked tasks pulled per deal.

Human touch. A note, completed task, synced email, or stage move performed by a named user, excluding anything executed by an automation and excluding the shared admin account's actions within 10 minutes of deal creation (lead-intake integration). Later admin-account work (2,770 actions) is shown in the shared-login finding but not credited to any agent.

Attempts. Notes + completed tasks + synced emails. Synced emails may include semi-automated sends from connected inboxes, so attempt counts for email-heavy agents (Elena, Ashley, Penny) are best read alongside their task/note columns. Calls made without logging leave no trace — all numbers are floors on real effort, but the same floor for everyone.

Speed-to-lead clock. Headline STL counts only time elapsed between 9am and 9pm America/Toronto, seven days a week — a lead created at 11pm and called at 9:15am scores 15 minutes, not 10 hours. The raw wall-clock median is shown in grey beneath each figure. "wh" columns = working hours.

Attribution. Deals are grouped by current owner; touches by other users still count toward the deal. Severity thresholds: first touch ≤1 working hour good / >6 critical; never-touched ≥10% warn, ≥25% critical; stale ≥50% warn, ≥85% critical.