Digital Empathy · the eight-week plan · Jul 13 – Sep 6, 2026

The Eight Weeks

The sprint calendar — owners, dates, and per-deliverable cards. Tap a name for your lane; tap a row for its card.

v7 · calendar first · tap a name for your view, tap a row for its card — annotate anything, comments come back to Robert

The calendar

View
⚑ = hard milestone · tap any line for its deliverable card · inside each card, More context opens the fuller story and where it's drawn from · tap a name for one person's view
Phase 1 · Ready + prove

W1 · Jul 13–19

WhatHand Derek the objective, not a plan: 25 signups ≈ $10K MRR in 8 weeks, warm segment first — his path, his roadmap.
Done meansDerek has the goal + the incentive frame, and his first-pass roadmap clock is running.
Depends onIncentive structure — Robert + Alie settle it this week.
More context

The retreat's delegation model in one move: hand the outcome and the KPI, never the path — ownership creates better plans than instructions do. Derek knows the VCI field reality better than anyone (he's been fielding the objections, including the recurring “I have Weave”), so the plan is his to write and the goal is the contract. The incentive piece exists because leadership decided real ownership deserves real upside; comp design is Alie's frame.

Drawn from: retreat — Prism & priorities session + eight-week-plan session (Jul 12); ruling Jul 13.

WhatŁukasz gets ONE goal — read-only Prism in as many practices as possible — plus the three pillars (revenue data · services data · hygiene) and the straight rate-limit/throughput question.
Done meansŁukasz can write his roadmap without guessing what matters.
More context

Why read-only first: the pain Prism removes is back-end edit friction, and read-only access is one build with two unlocks — VCI onboarding stops being manual, and copywriting discovery stops being “pulling teeth to get the services list.” The rate-limit question is asked straight because the honest throughput answer decides how fast anything downstream can scale. “Go explore” was explicitly named as failed direction — one goal and three pillars are the fix.

Drawn from: retreat — Prism & priorities session + claims review (Jul 11–12).

WhatFour commitments land the same day: Eve MVP (done, or a delineated list with timelines) · first-pass team-test criteria (start from the settled 10-case anchor) · the Neuron edges/unknowns list into Alie's hands · the check-in itself (roadmap walk, VCI incentives, Marcy's working surface).
Done meansAlie can write the customer-pilot criteria — the edges list is what gates her.
Depends onAlie's definition-of-done doc (Eve MVP is graded against it).
More context

This is the week's load-bearing wall — the check-in moved from Monday to Wednesday precisely so these could converge. The edges list matters because Alie's standing position is that confidence must come from data, not assurances: she writes the customer-pilot test criteria only once she knows where Neuron excels and where it's less sure. The 10-case test size was settled at dinner on Jul 12 (30 was considered and rejected) — starting there avoids re-litigating a decided number.

Drawn from: retreat — Prism & priorities session (Jul 12); Robert's markup (Jul 13–14).

WhatNotify-with-opt-out (not opt-in) to the first 20+ sites — the sprint's first domino.
Done meansSent — and no client ever learns about their migration after the fact.
Depends onRobert's red-pen → Alie comfortable. Her gate holds the claims line: nothing in the email we can't already prove.
More context

The posture was decided at the retreat: “we've made this big investment… if for any reason you don't want us to do it, let us know by this date” — notification with a real exit, not a permission request. The discipline underneath: communication confidence must never outrun pilot-data confidence, so the email carries what we verify (item-for-item inventory match before any switch) and promises measurements, never outcomes. Alie's comfort is the formal gate on the send.

Drawn from: retreat — eight-week-plan session (Jul 12); Robert's markup (Jul 13); draft in `drafts/migration-notify-draft.md`.

WhatThe pilot cohort comes from the sheet of practices we've been in pilot conversations with for months. Robert works it with Arby's help — the last 10% these conversations have been waiting on.
Done meansFirst deliverable: pilot outreach email copy with Alie's sign-off — accurate, nothing over-claimed, and not framed so cautiously it invites an easy no. Then outreach out, pilots scheduling.
DueEmail copy this week; cohort moving by W2.
More context

The sheet predates this plan — practices that already said yes in principle, where only the final scheduling push never happened. That's why the cohort decision (Jul 14) was to work this sheet rather than recruit fresh: warm beats cold, and finishing an open loop is the plan's own medicine. The email's two-sided bar is deliberate: Alie's claims gate on one side, and on the other, framing where the only decision offered is this week or later — never yes-or-no.

Drawn from: retreat — claims review (Jul 11); Robert's ruling (Jul 14); draft in `drafts/pilot-outreach-draft.md`.

WhatEvery AVMA lead into Salesforce with an owner. Not all of it is VCI, so the sheet isn't Derek's lane — Robert owns it with Arby's help, and VCI-interest leads route to Derek warm.
Done meansEvery lead logged with an owner and a first touch made; VCI-interest leads in Derek's hands.
More context

This one almost got lost: the commitment was made the weekend of the conference, then fell out when the working docs were rebuilt after the retreat — the completeness audit caught it. Ownership was settled Jul 14: the sheet spans more than VCI, so it can't just be Derek's; Robert owns it with Arby doing the logging-and-drafting labor. Speed still matters — conference warmth decays — but the hard SLA came off because the clock had already run past it; the bar now is simply: every lead owned and touched.

Drawn from: war-room working docs (Jul 12); Robert's ruling (Jul 14).

WhatWhere Arby is today, real drafts, access — plus the two lead claims and the response-type structure, so the bar work starts from substance.
Done meansJess's W2 bar + gap analysis has everything it needs — including her definition of the vet persona and needs Arby must communicate to.
Fails loudlyPackage late this week = her W2 analysis starves.
More context

The original idea was a four-month audit of past support questions; it was replaced (Jul 13) with a faster path: hand Jess the live package, she writes the bar. Her unique standing is the reason this lane exists at all — she's a vet, she knows how veterinary professionals actually digest information, and she was asked to hold the quality bar without ever lowering it. The persona/needs definition (Jul 14 note) is the concrete first artifact: it tells Arby who he's talking to before anyone grades how he says it.

Drawn from: retreat — support-materials session (Jul 12); rulings Jul 13–14.

What~100 sites have no DNS cut-over information — the wall in front of “all sites migrated.” Matt holds the list; Arby's assist mines what we hold, names the per-site ask, chases registrars/hosts.
Done meansBurn-down motion visible — Matt, Marcy + team working the list, count dropping against weekly targets (set in W2).
Depends onMatt's list handed over this week.
More context

DNS is the part of a migration we don't fully control — it depends on who holds each domain, which is exactly why it's the schedule risk. The Arby assist has three jobs per site: mine everything we already hold (old emails, registrar records, past tickets), name the single specific ask that unblocks that site, and draft the chase. Weekly targets rather than one end-date cliff (Robert's call, Jul 14) — roughly a quarter of the list per week, exact split set with Matt when the list lands.

Drawn from: Robert's markup (Jul 13–14); Matt's list (pending).

WhatDerek: his first-pass plan to 25 — his lane, his path; the goal is the contract. Łukasz: the Prism 8-week roadmap — what revenue data is actually pullable, services-data reality, his practice-count target, and the honest rate-limit/throughput answer.
Done meansBoth on paper by end of week; each sets the number their own lane is measured by — then each gets a quality walk at the Friday four-liner (the objective stays theirs; the plan gets eyes).
More context

These two roadmaps ARE the upstream planning for their lanes — by design. The plan hands objectives, not paths, so the calendar deliberately doesn't prescribe how Derek sells or what Łukasz builds first. The quality walk exists so “his lane, his plan” never quietly becomes “nobody read the plan”: Friday's four-liner is where each roadmap gets real eyes without taking the pen out of the owner's hand.

Drawn from: retreat — Prism & priorities session (Jul 12); governance research digest.

WhatThree seed sources for the proactive practice: upgrades we fumbled (Marcy's outreach list) · the ads-cancellation watch · mining MailChimp opens we already hold.
Done meansLists exist and the first saves are in motion before Marcy's back full-time in W2.
More context

The ads-cancellation watch is the sharpest early-warning signal we have: ads cancellations predict website cancellations about two months out, and the last four months saw an unusual spike — the canary may already be singing. Fumbled upgrades are the warmest possible save (they already asked for something). MailChimp opens are the cheapest signal (we already hold the data; someone opening emails about a service is raising a hand). All three exist so the practice starts on real signal, not cold guessing.

Drawn from: retreat — Prism & priorities session (Jul 12); fleet census (Jul 12); war-room working docs.

WhatThe registry: every initiative gets an owner, a burn-down clock, and a loud failure mode — visible to everyone. And the months-old SEO ticket backlog finally gets an owner — worked as a migration play: “we'll fix it with the move.”
Done meansRegistry live with every lane on it; SEO tickets owned, each one paired with a Neuron switchover offer.
More context

The registry is the cure for the named disease — initiatives that start with enthusiasm and end silently — and it also closes a gap Alie flagged: team projects and leadership work sharing no common map. The SEO two-birds framing is Robert's (Jul 14): a client with a months-old SEO ticket is a client with a reason to say yes to the new foundation, so the backlog isn't just debt to clear — it's a warm list for the switchover. Fix the ticket, make the move, one conversation.

Drawn from: retreat synthesis; Robert's markup (Jul 14).

W2 · Jul 20–26

WhatJess writes the bar: what a 10/10 client communication sounds like to a veterinary professional, where our current output sits, and the named spaces between.
Done meansThe bar is in writing and Robert can implement against named gaps — drafts get re-scored from here on.
Depends onThe W1 communication package.
More context

The bar has three signals — curiosity, understanding, expertise — named at the retreat as what every client touch should carry. Jess holds the clinician lens (how a vet hears it), Alie the communication lens (how DE says it); they hold the bar together, and Robert folds the findings into Arby. The loop only works if the gap is named, not felt — “this draft is a 6, here's the space between it and a 10” is what makes improvement implementable.

Drawn from: retreat — fires-now + support-materials sessions (Jul 12, with Jess); rulings Jul 13–14.

WhatThe pilot cohort (from the existing pilot sheet) goes through the test series defined Wednesday: page speed, bot traffic, before/after SEO.
Done meansAlie's confidence gate is satisfied with data, not assurances — that's what unlocks the migration timeline.
Depends onWednesday's test criteria + the pilot outreach landing.
More context

This is the hinge of the whole migration lane. Twenty pilots through a defined test series is Alie's stated gate for trusting the timeline — her position from the retreat: she can't take anyone's word that the migration is safe until there's data that shows it. The support team runs pilots too, which is what proves lane 2 (support readiness) at the same time. Everything after W2 — scaling, the all-sites date, the proof package — stands on what these twenty show.

Drawn from: retreat — eight-week-plan session (Jul 12); Robert's ruling on the cohort (Jul 14).

WhatThe segments doc — who we're for, sliced for action.
Done meansThe W3 channel decision is made by this doc, not by defaulting to email.
More context

There was a real disagreement here at the retreat — start emailing the 20k list now, or let segments decide the channel — and it resolved to Alie's position: segments dictate channel. This doc is where that resolution becomes real. It also sets the design constraint for whatever channel wins: open with self-evident help, give them nothing to sell — de-risking is the conversion lever for practice owners.

Drawn from: retreat — eight-week-plan session + claims review (Jul 11–12).

WhatMarcy picks up the seed lists and runs the outreach; the tighten-the-loop conversation happens (her wish list for what would let Arby up-level the work).
Done meansOutreach running; her wish list captured and feeding Arby improvements.
More context

Marcy's missed-opportunities work was named at the retreat as the existing middle ground between heroic last-minute saves and the future-state fix — the proactive practice grows out of what she already does, it doesn't replace it. The wish-list conversation matters because she's the operator: what would make this work better is a question for the person doing it, and her answers become Arby's build list.

Drawn from: retreat — fires-now + Prism sessions (Jul 12).

WhatThe practice gets designed for real, on paper: Marcy + Robert draft it with Arby's help; Alie + Jess pass it.
Done meansA one-page spec exists and is passed before W3 — the practice starts its weekly rhythm from a design, not a vibe.
Fails loudlyW3 starts without a written spec = we're winging the exact thing this plan exists to fix.
More context

This row exists because of an honest admission (Robert, Jul 14): “we don't even totally know what needs to happen there.” The lane had a goal, a loop, and seed data — but no designed practice. Rather than pretend, the design itself is now the deliverable: one page answering what goes out, to whom, on what cadence, which parts Arby carries, and how we'll know it's working. The W1–2 seed work feeds it real observations, so the spec is written from evidence, not theory.

Drawn from: Robert's lifecycle review (Jul 14); retreat — fires-now session (Jul 12).

WhatTurn the ~100-site wall into weekly goals, not one cliff — shape: roughly a quarter of the list per week, exact split set with Matt when the list lands.
Done meansA weekly number exists, and each Friday four-liner reports against it.
More context

Checkpoint thinking, applied the same way VCI got its Aug-9 midpoint (Robert's note, Jul 14): a single “all resolved by W6” date hides slippage until it's too late to react. Weekly targets make the burn-down visible while there's still time to change approach — and they give the Sandbox gate something concrete to call Go/Hold on.

Drawn from: Robert's markup (Jul 14); governance research digest.

WhatConcrete tooling progress with Jess on the editorial world-model — the commitment is real progress she can feel inside the fortnight, not a someday.
Done meansSomething working in her hands, and a date set on her decomposition doc.
More context

This is a commitment made directly to Jess and it's tracked here so it can't quietly slip. The shape: decompose what she carries in her head into pieces tooling can hold, so the editorial load stops depending on one person's ceiling. It pairs with a standing protection written into this plan's go-to-market lane: we don't sell more copywriting capacity with Jess's bandwidth as the price.

Drawn from: retreat commitments (Jul 12); this plan's lane 7 constraint.

WhatRebuild the outbound system on the ~20k-practice database, enriched with behavioral signals: sites frozen in time (Wayback), vendor identity, VHMA sentiment.
Done meansA segmented, signal-rich list ready for the W3 channel call.
More context

The database already exists (~20k practice emails, pulled about six months ago) — the rebuild is about signal, not size. Behavioral prospecting was one of the retreat's sharper ideas: a site that hasn't changed in years (visible via Wayback), the vendor a practice currently uses, and how their community talks — each is a reason a specific practice might move, which is what turns a list into segments Alie's doc can actually use.

Drawn from: retreat — eight-week-plan session (Jul 12); fleet census.

WhatThe rhythm goes live: Friday four-liners as pre-read, registry walked, gates called.
More context

First live turn of the operating rhythm (see the Rhythm section): written four-liners Friday, Monday for exceptions only, Sandbox as the gate that can actually say Hold or Kill. Arby carries the labor in #sprint — drafting, nudging, flagging — so the recap costs owners minutes, not an afternoon.

Drawn from: governance research digest; Robert's markup (Jul 14).

Phase 2 · Turn outward

W3 · Jul 27 – Aug 2

WhatFirst campaign out on whatever channel the segments doc says. Standing constraint: open with self-evident help — give them nothing to sell.
Done meansCampaign live, measured, feeding pipeline.
Depends onSegments doc (Fri 24) + practice-database rebuild.
More context

Channel ownership after the segments doc is a named open question, on purpose — the doc decides, then an owner is named at the W3 Sandbox. Deeper design work (audience, offer, sequences) belongs to whoever takes the channel; this calendar holds the date and the constraint, not the playbook.

Drawn from: retreat — eight-week-plan session (Jul 12).

WhatThe practice starts its weekly rhythm from the passed spec. Plus the escalation middle rung — one-pagers so support isn't stuck between “ask Robert” and nothing.
Done meansv1 cadence running; one-pagers in support's hands.
More context

The escalation gap was named plainly at the retreat: today the ladder is “a call with Robert” or nothing, which makes him the bottleneck and leaves support stranded on anything in between. One-pagers are the cheap middle rung — enough to resolve the common hard cases without an escalation, and they fit naturally in Alie's support-readiness lane.

Drawn from: retreat — fires-now session (Jul 12, with Jess).

WhatWith pilots proven, migrations run at volume. 1:1 resets start — thoughtful but time-economical, leadership messages aligned before any conversation happens.
More context

The resets replace the skipped mid-year review cycle with something lighter and more honest: expectations per person, reset in one good conversation each. The one hard rule, set at the retreat: Robert and Alie align the message before the conversations start — nobody hears two versions of their own expectations.

Drawn from: retreat — eight-week-plan session (Jul 12).

W4 · Aug 3–9

WhatThe first scheduled gate we mean: at 10 signups we're on pace; short of it, Sandbox diagnoses — motion, segment, or offer — and adjusts.
Fails loudlyA midpoint miss with no diagnosis attached.
More context

The midpoint exists because an eight-week goal with no interior checkpoint can fail invisibly for six weeks. “Diagnosis, not punishment” is load-bearing: the useful question at 7 signups isn't “whose fault” but “which part of the motion is wrong — who we're calling, what we're offering, or how.” The gate framework (Go / Hold / Kill / Recycle) gets its first real test here.

Drawn from: rulings Jul 13; governance research digest.

WhatThe loop turns weekly: Marcy runs it, Alie + Jess score against the bar, Robert folds findings into Arby.
Fails loudlyEval findings piling up unimplemented.
More context

The full loop, running: sense (Marcy's outreach + the watch lists) → judge (Alie + Jess against the written bar) → improve (Robert implements into Arby) → the next week's touches are better. The fails-loudly line targets the known failure mode: feedback that gets collected and never built.

Drawn from: rulings Jul 13; retreat — fires-now session.

Phase 3 · Scale + finish

W5 · Aug 10–16

WhatThe proven standard applied across the fleet, DNS burn-down feeding the queue, ops metrics logging all the way.
More context

By this week the open questions should all be closed: standard proven (W2), clients notified (rolling), DNS burning down on weekly targets. If any of those is still open, this week inherits the slip — which is exactly what the weekly gates upstream exist to catch early.

Drawn from: the migration lane's dependency chain.

WhatWhatever the midpoint diagnosis said, applied: the motion that works, run harder.
More context

Post-midpoint, the lane's job changes from “find the motion” to “run the motion” — and the documented, repeatable version of it is itself a Sep-6 deliverable, so what's learned scaling is written down as it happens.

Drawn from: the VCI lane contract.

WhatThe materials that cleared Jess + Alie's bar stop being drafts: they ship inside real client replies, with honest-odds framing throughout.
More context

The test of the bar work is whether real clients get better answers — not whether a binder exists. “Honest odds” is the standing register: here's the playbook, here's what improves the odds — the way a good doctor talks — never promised outcomes.

Drawn from: retreat — support-materials session (Jul 12); the no-promised-outcomes ruling.

W6 · Aug 17–23

Done meansEvery site we manage on Neuron, zero surprised clients along the way.
Depends onThe DNS burn-down clearing the ~100-site wall on its weekly targets.
More context

The six-week line set at the retreat. The honest risk to it was named from day one: the ~100 sites where the DNS cut-over isn't in our hands. That's why the burn-down runs on weekly targets from W2 — if the wall isn't falling fast enough, Sandbox sees it with weeks in hand, not days.

Drawn from: retreat — eight-week-plan session (Jul 12); Robert's markup (Jul 13–14).

Phase 4 · Prove + package

W7 · Aug 24–30

WhatSpeed, bot traffic, SEO before/after across the fleet — plus the operational record captured all along (fleet time, per-site time, training time). Pure Neuron: the proof-of-concept for future marketing, and the thing that kept the fleet migration organized.
Done meansThe migration claim is provable with receipts — ours, from our own fleet.
More context

Two jobs, one package. Inward: the operational record is what keeps a hundred-site migration honest while it's happening. Outward: this becomes the case-study foundation for whoever we make the migration case to next — which is why the record includes how long things actually took, not just how the sites performed. It has nothing to do with VCI; it's the Neuron story, told in our own measured numbers.

Drawn from: Robert's markup (Jul 14); the migration lane's ops-metrics commitment.

WhatThe last push to 25 — pipeline built in W3–5 closing now.
More context

Separate line from the proof package on purpose: VCI's finish is a sales sprint, Neuron's finish is an evidence package. Related stories, different lanes, different owners.

Drawn from: Robert's markup (Jul 14).

W8 · Aug 31 – Sep 6

Done means25 paying practices, with the repeatable motion documented.
More context

The number was set at the retreat: 25 practices at VCI pricing is roughly $10K in monthly recurring revenue — real money, and more importantly a documented motion that can be run again. The goal is the contract; the path was Derek's to design from W1.

Drawn from: retreat — eight-week-plan session (Jul 12).

Done meansRead-only Prism at the practice-count Łukasz's own W1 roadmap set, and the revenue source-of-truth question answered.
More context

Related to VCI — Prism kills VCI's onboarding friction — but a separate goal with a separate owner, kept on its own line so neither lane's finish hides the other's. The target is deliberately Łukasz's own number: he sets it in W1 with honest knowledge of rate limits and data reality, then he's measured against it.

Drawn from: retreat — Prism & priorities session (Jul 12); Robert's markup (Jul 14).

WhatThe eight weeks get the same treatment as every initiative: a post-mortem — what we learned, what we keep.
More context

The plan eats its own cooking: every initiative that completes gets a lightweight post-mortem, so the sprint itself closes with one. The question isn't “did we hit everything” — it's which parts of the machinery (the rhythm, the gates, the registry, Arby's labor) earned a permanent place.

Drawn from: the plan's own operating rules.

How a project runs here — the protocol & process (tap to expand)

When to run a room: the moment work needs other people’s time — or wants to ship to the team or a customer — it gets a room (2–4 people). Until then, explore freely.

  1. Before. Name the question in one line · pick the 2–4 people who’ll actually decide · start Granola when you start talking. No pre-doc.
  2. In the room. Think out loud · surface disagreement here, not after · demos beat descriptions.
  3. The gate. Before anyone leaves, say it out loud: what we’re doing · who owns it · when it lands. Can’t say all three? Say what’s missing, who gets it, when we reconvene.
  4. After. Eve distills the transcript · the consensus goes to #sprint · the owner builds against what the room said. Friday four-liner keeps the clock honest (Arby drafts, owner approves).
The consensus block (copy into #sprint on the way out):
What we’re doing: ____ · Who owns it: ____ (one name) · When it lands: ____ (a date)

Fine print: one owner per project (two owners is zero owners) · a date is a date — if it moves, that’s said in #sprint, not discovered · reconvene beats drift. Each calendar row above opens the same way — tap a deliverable for its card: what it is · what done means · when it’s due · what it depends on.

The lanes (owner's reference)

  1. Neuron migration — Robert. Goal: every site on Neuron by ~Aug 23; clients notified honestly; support confident.
    what must be true
    • Definition-of-done sprint: the standard, proven on 20 pilots (speed, bot traffic, before/after SEO)
    • Pilot cohort = the existing pilot sheet; Robert re-engages with Arby's help — first deliverable: outreach email copy with Alie's sign-off
    • Public claims ride behind pilot data — nothing goes in an email we haven't measured. Alie holds that gate.
    • Notify email with opt-out window → first 20+ sites (Wed)
    • Migration ops metrics captured from day one — fleet time, per-site time, training time — part of the proof package
    • DNS gap: Matt holds the ~100-site list; Arby assist built so Matt, Marcy + team burn it down against weekly targets
    • Done: all sites migrated · zero surprised clients · proof package exists
    • Fails loudly: a client learns after the fact · pilot tests undefined at end of W1
  2. Support readiness — Alie. Goal: support team confident on migrated sites, tools to match.
    what must be true
    • W1: Alie onboarded to make UI changes herself — then the team trained on it (the real post-conference work)
    • Support UI → team (W1) · team proven on pilots (W2) · escalation one-pagers (W3–4)
    • Input she's owed: the Arby support-UI handoff doc from Robert (where it's strong, where it's weak, decisions made, principles)
    • Done: support runs on migrated sites without escalation spikes
  3. VCI — Derek, with incentives. Goal: 25 signups ≈ $10K MRR by Sep 6.
    what must be true
    • Roadmap W1 — his lane, his plan; the goal is the contract
    • Incentive structure: Robert + Alie settle this week
    • Warm segment first (ads + call-tracking clients, plus VCI-interest AVMA leads routed from Robert) · objections synthesized thematically
    • Midpoint: 10 by Aug 9
    • Done: 25 live + repeatable motion documented · Fails loudly: no W1 roadmap · midpoint miss with no diagnosis
  4. Prism — Łukasz. Goal: read-only practice data from as many practices as possible.
    what must be true
    • Roadmap W1: what revenue data is actually pullable · services-data quality reality · the honest rate-limit/throughput answer
    • Three pillars: revenue data · services data · hygiene · progress visible weekly
    • Done: his roadmap's practice-count hit + revenue source-of-truth answered · Fails loudly: a silent week · an unraised blocker
  5. Proactive communication — the new practice. Goal: every client touch signals curiosity, understanding, expertise — before the client asks.
    what must be true
    • The loop: Marcy runs it · Alie + Jess hold the bar (clinician lens + communication lens) · Robert implements improvements into Arby
    • Human + agent, complementary — we never ship Arby and say “Arby's got it covered.”
    • W1–2 seed: fumbled upgrades → Marcy's list · ads-cancellation watch (~2-month churn early-warning; recent spike) · MailChimp-opens mining
    • W2: the practice gets designed for real — a one-page spec (what we send · to whom · cadence · Arby's parts · how measured); Marcy + Robert draft with Arby's help, Alie + Jess pass it
    • W3+: runs weekly, measured · working surface: start in Slack with Arby; purpose-built UI decided after real use
    • Marcy runs this sprint's practice; the long-term shape of the relationship-owner role stays an open question — revisit with W2 results in hand
    • Done: named cadence running without heroics + one saved at-risk account we can point to
    • Fails loudly: W3 starts without a written spec · stays a value instead of a calendar · eval findings pile up unimplemented
  6. The communication bar & materials — Jess (with Alie). Goal: a written bar for how DE communicates with veterinary professionals — and materials that clear it.
    what must be true
    • This week: Robert hands over the package (Arby today, real drafts, access, the lead claims + response-type structure)
    • From the package, Jess also defines the vet persona + needs Arby must communicate to — she's the one who knows how veterinary professionals digest information
    • Next week: Jess returns the bar + gap analysis (what 10/10 sounds like · where we sit · named spaces)
    • Then: Robert implements against gaps → drafts re-scored → materials emerge from the closing loop
    • Robert communicates reasoning before Arby changes ship — the bar-holders are never surprised by the thing they're grading
    • Honest-odds framing throughout (playbook + what improves the odds; no promised outcomes)
    • Done: bar in writing · gap shrinking on re-scores · top questions answered vet-credibly · Fails loudly: package late this week · analysis missing next week
  7. Go-to-market — Alie (segments), then channel owners. Goal: new-practice marketing — segments → channel → first campaign — live by W3–4.
    what must be true
    • Segmentation doc end W2 · channel decided by segments, not default
    • Outbound system rebuilt W2 — practice database (~20k) enriched with behavioral signals (frozen sites via Wayback, vendor identity, VHMA sentiment)
    • Constraint: open with self-evident help, nothing to sell
    • We don't spend Jess to sell copy — no selling copywriting capacity priced in her bandwidth
    • Done: first campaign out, measured, feeding pipeline
  8. Team ops — Robert. Goal: the machinery that makes initiatives finish — and a team that never queues on one person.
    what must be true
    • W1: Eve MVP · blockers cleared · SEO backlog owned (ticket sites → switchover candidates) · initiative registry up: owner + burn-down clock + loud failure, visible to all — closing a gap Alie named: team projects and leadership work sharing no map
    • W2: weekly Sandbox sprint recap real · blockers sourced from the team directly
    • W3–4: 1:1 resets · frontier tooling handed with bounded asks + success signals, never "go explore"
    • Standing: completed initiatives get a lightweight post-mortem — what we learned, what we keep — prompted automatically
    • Done: registry live with every lane · two Sandbox recaps done · blockers cleared and raised in the open
    • Fails loudly: an initiative ends silently

The sprint rhythm — governance without theater

Shape drawn from how fast-moving companies actually run cycles — Basecamp's Shape Up, Linear's method, Amazon's weekly business reviews, Stripe/GitLab written-first culture, demo-day practice, and the empirical standup research — filtered to 13 people and 8 lanes. The headline: we already have the right number of meetings; the upgrade is what happens inside them.

1 · The Friday four-liner

Every lane owner, in Slack, ~10 minutes: Shipped · Next · Blocked · Clock (on-track / behind — and if behind, what changes). Same format every week; link real artifacts, don't describe them.

2 · Monday = exceptions only

The four-liners are the pre-read. Sync time goes to what's contested, blocked, or being decided — never to re-reading what's written.

3 · Sandbox = the gate

Leadership walks the registry weekly: any lane whose fails-loudly condition fired gets an explicit call — Go / Hold / Kill / Recycle — against conditions set before the sprint. A gate that always says Go is theater; Aug 9 is the first test that we mean it.

4 · Demos over status

Running work gets shown live — real artifact, no slides, opt-in. Mandatory demos make people optimize for the demo instead of the work.

5 · No daily standup

At 13 people and 8 independent lanes, standups become reporting-to-the-manager theater (the research is blunt). Day-to-day unblocking lives in lane channels, async.

6 · Arby + Eve carry the labor

People judge; Arby labors. Home: the #sprint channel, where everyone lives. Arby drafts each owner's Friday four-liner from the week's real artifacts (owner edits and approves — never auto-posted), keeps the registry current, and flags any fired fails-loudly line before Sandbox walks it. Eve automates the nudges, clock states, and post-mortem prompts.

This rhythm absorbs the Hub weekly recap — same job, sharper format. Alie confirms the swap before anything is retired.

How we work this plan

Owner · timeline · loud failure

Every initiative, all three. Quiet rows are the signal. Finished initiatives get a lightweight post-mortem: what we learned, what we keep.

Objectives, not plans

You own the path; the goal is the contract. Early asks are strength.

Prove it before plans change

Bounded problems, proven capability, no unmeasured claims.

The client-touch bar

Curiosity, understanding, expertise — and we know where our knowledge ends.

No promised outcomes

The playbook and the odds, like a good doctor.

Open items being worked

Living document. Owners update their lane weekly at the Sandbox recap. Annotate anything on this page — comments come straight back to Robert. Companion: Alie's one-screen decision sheet, refreshed with the Friday rhythm.