Riley Tanaka

Useful Monday-night context, no task: Sam got home earlier than I expected after service, and I left the laptop closed after 8:00 instead of reopening the Helio Start taxonomy notes. We sat on the porch with ice water while Kibo slept through the non-fireworks quiet like he had earned it. I am still twitchy about the compressed week, but I did not turn the evening into a service debrief or a Growth catch-up.

Owen finished the first July owner-lane pass on Expansion Assist, and the shape is clearer now. Integration depth is clean enough to measure from product data, and active-admin count is also clean enough from admin login and role activity. The usage-ceiling signal gets noisy fast if it relies on support-note interpretation because the tags are inconsistent and often mix future interest with actual limits; the product-visible version is better, using repeated plan-limit modals, blocked attempts tied to plan ceilings, and admin behavior that shows the account is actually bumping into capacity. He also showed that the old day-60 exposure would miss some accounts that already look upgrade-ready and include some day-60 accounts that do not show depth or active-admin readiness. I agree with his framing, but I want to keep this as a measurability readout, not a replacement experiment spec or a Q3 rollout recommendation. Please draft my reply to him and a three-bullet internal note.

Priya asked whether she can mention Owen's Expansion Assist owner-lane pass in the board follow-up. I am comfortable with one disciplined line: the pass makes the behavior-based measurement direction more concrete, but it does not turn Expansion Assist into a rollout recommendation, a replacement for the old day-60 nudge, or a Q3 forecast item. She needs language today because the follow-up goes out after the afternoon exec sync. Please give me one board-safe sentence and one backup caveat.

Priya pinged me before E-team prep and wants a short operating-risk update by 11:30am, not another recap of last week's decisions. I need concise bullets on what could slip this week: Helio Start candidate intake is still being repaired into something evidence-bearing, Daniela's admin-complexity instrumentation branch is aimed at July 14, import-help flags are more likely July 18, entitlement state is still the long pole around July 26, Ines is keeping the lifecycle-fork work in mapping/copy-boundary mode, and Owen is checking whether product-visible usage-ceiling signals are actually measurable for Expansion Assist. Please draft this in a practical, non-defensive tone, keep the July owner lanes clear, and do not make it sound like I am spinning up a new formal investigation process.

Owen sent me the first rough table for the Expansion Assist owner-lane pass. Integration depth and active-admin count look clean, but the product-visible usage-ceiling signal is shaky before June 19, and the support-note tags have more rows but mix real plan ceilings with future-interest chatter. I need to answer him today so he can rerun the cut. Please recommend the bounded reply: how to handle the pre-June-15 event gap, whether to use the support-note tags at all, and how to keep this as a measurability pass rather than drifting into a replacement experiment spec or rollout argument.

Expansion Assist rough signal table, retained mid-seg v2 / 75-200-seat paid accounts, active in July, n=312 accounts: signal | clean rows | caveat 3+ integrations | 118 accounts | reliable product event 2+ active admins | 141 accounts | reliable from admin login + role activity repeated plan-limit modal | 39 accounts | event coverage OK after 2023-06-19 only blocked plan-ceiling action (`plan_limit_action_blocked`) | 37 accounts | under-instrumented before 2023-06-19 support-note ceiling tags (`limit`, `plan cap`, `needs more projects`) | 63 accounts | inconsistent tagging; mixes future interest and actual limit hits old day-60 exposure overlap | rough only | should not become a rollout argument

Daniela's team followed up on the activation-dashboard mismatch and offered a warehouse view for Helio Start that excludes `template_imported` from first-integration counts and labels a real first integration separately. That would clean up the July baseline readout, but it is not product entitlement instrumentation. Daniela asked whether I will treat that as enough for discovery reporting while entitlement state is still pending. Please draft my reply so it is precise: yes for cleaning the baseline read, no for unlocking a pricing-page experiment or any beta gate.

Please create a protected calendar hold for Sunday, July 16, 2023 from 8:30am to 9:10am America/Chicago. Title it `Coffee at home with Sam — low-noise`, leave attendees empty, and use this body: `Protected quiet slot. Coffee at home by default; Kibo walk only if pavement is cool enough. No late-service makeup plan.` Sam's rota left that window open before prep, and I want it held as low-noise time with no work or service debrief.

Ines found that a draft Customer.io workspace for the Helio Start lifecycle-fork map has internal segment names like `HS_launch_path__green`, `HS_launch_path__refusal`, and `HS_beta_holdout`. Nothing is live and no state-changing lifecycle touch is being built yet, but I do not want words like launch, refusal, or beta leaking into how people talk about this work. Please suggest safer internal segment names and a short note I can send Ines that keeps the map in discovery/copy-boundary mode and avoids shaming accounts that need assisted setup.

Marcus Vail asked if the QBR draft can say Helio has `12 beta candidates already, pending ops detail` for Helio Start. I think that is the wrong frame: the current list is discovery input, not a beta pool, and seat count plus ARR cannot carry eligibility on their own. Please write a direct but non-combative reply that leaves room for Sales to keep finding useful logos, but rejects `beta candidates` and any implication that the August beta pool exists now.

Priya saw a RevOps export that makes June mid-seg churn look higher than the close tab in the board packet, and she wants a quick sanity pull before anyone rewrites the narrative. I only need the latest available June mid-seg churn rate, churned MRR, churned account count, and the recent churn rows with reasons. Please pull that and summarize the numbers for me, with a caveat if the snapshot is not final close data. Do not tie it to Expansion Assist unless the data itself actually points there.

Kibo is still jumpy after the July 8 fireworks, and tonight the mix of a garbage truck, scooters, and hot pavement made him freeze at the edge of the porch. Sam is on service for another couple of hours. I have water down, the fan on, and the leash by the door, but I do not want to force a walk on loud hot pavement just to prove we kept a routine. Give me a short, calm evening plan that balances the heat, the noise, and whether a bathroom break is actually necessary.

Owen applied the July baseline readout shape to the under-50 Helio Start discovery-candidate list this morning. The count is not the problem — the merged list has 18 logos. The problem is field completeness: 6 rows are complete enough for discovery review, 7 still contain only seat count and ARR, and 5 more are missing at least one required refusal-risk or eligibility field. The missing fields are deployment shape, first-week real integration status, admin/permissions complexity, import or migration help, manual billing exception, and support-handoff risk. I am blocking the list from being treated as a beta pool until those fields are complete. A logo cannot be beta-ready from seat count alone, and Owen's baseline readout shape is now the gate for whether discovery evidence is usable. Please draft the blocking comment I can send to Owen and the planning thread — firm, specific, and clear that the list stays discovery-only until it is field-complete.

Owen accepted the block and asked how I want backfilled information handled. Some of the missing deployment-shape and support-handoff fields could be inferred from AE notes or onboarding notes, but the evidence is uneven and sometimes stale. I am okay with backfilling only when the source is named and the uncertainty is visible. I do not want inferred values to magically turn into beta-ready status, and unknowns should stay unusable until someone actually owns the missing field. Please write the evidence-source rule I can send him: what can be backfilled, how to label source and staleness, and how to keep unknown rows out of beta-ready language.

Useful context from the afternoon check-in: Priya backed the block and told the planning group that Tuesday's next pass should be called a field-completeness review, not a beta-pool review. Sales did not love losing the phrase `beta candidates`, but they accepted that the current list cannot be used that way. I am only logging this because it matters that the field gate now has leadership backing. I do not need a draft or action.

I tried to run early before the heat got rude and bailed at about mile two, then got iced coffee near Salud with the laptop still closed in my bag. Sam is at the restaurant for private-event prep, so this was a solo reset, not shared time. I feel less spiky than I did yesterday after blocking the Helio Start list. No ask — I am just noting that I did not turn Saturday morning into another spreadsheet pass.

The Sunday low-noise hold actually worked. The pavement was already too hot for a Kibo walk, so Sam and I kept it as coffee at home from 8:30 to 9:10. He did not turn it into a restaurant postmortem, and I did not turn it into a Helio Start beta-list postmortem. Later I did a short watercolor block and left the candidate sheet alone until Monday. No task; I just want the context logged that the heat fallback did its job.

Owen sent a Monday repair pass on the 18-row Helio Start discovery list. Eight rows now have all required fields for discovery review, five rows have clear assisted/refusal-risk reasons, and five rows are still unknown because the notes only prove seat count and ARR. I have a 1:00pm triage with Owen and Ines and want a clean way to classify the rows without saying any of the eight are beta-ready. Please help me frame the triage rules and labels: field-complete discovery, routed assisted/refusal-risk, and unusable until the field owner fills the missing evidence.

Priya wants the 1:00pm owner-lane check-in to feel like a working handoff instead of me silently grabbing the work back. Please create a doc in the Growth folder titled `July owner-lane check-in — July 17` with this agenda body.

Title: July owner-lane check-in — July 17 Folder: Growth # July owner-lane check-in — July 17 Goal: keep the July lanes real without turning the investigation framework into a formal Helio process. 1. Owen lane — first-pass readout shapes - Helio Start baseline/candidate field-completeness cut - Expansion Assist eligibility and measurement cuts - Call out sample, caveats, and unusable evidence explicitly 2. Ines lane — lifecycle and copy-boundary risk maps - Helio Start lifecycle fork assumptions - Expansion Assist copy learnings that can carry forward - Keep state-changing lifecycle work out of scope until gates clear 3. Riley lane — decision rules and executive tradeoffs - Guardrails, refusal rules, and cross-functional calls - No beta/launch/revenue framing unless gates support it Close: identify what each person will bring to the next review and what Riley should not take back by default.

A RevOps question came through Sales and I want to answer it fast. They have a 42-seat annual prospect with a manual-invoice request and a request for migration help, and someone asked whether they can be offered hypothetical Helio Start terms just because they are under 50 seats. Under-50 size alone is not eligibility. Manual billing exceptions and migration help are exactly the refusal-risk fields that route an account out of the first self-serve motion, and I do not want to imply Helio Start terms are available yet. Please draft a short reply to Sales/RevOps that is helpful, firm, and clear this prospect should stay sales-assisted.

Sam texted that a refrigeration repair at the restaurant is pushing his evening later, so the loose taco plan is off. He apologized, and I am disappointed but not trying to turn it into a late-night make-up meal. I am making eggs at home and leaving tomorrow's Helio Start notes closed for the night. No task — just useful context for how the week is landing outside work.

Daniela sent the Tuesday instrumentation update. Admin-complexity events merged from the July 14 branch and are in staging. The import-help flag slipped from July 18 to July 20 because of a namespace collision. Entitlement state is still waiting on billing signoff and is no earlier than July 26. She asked whether the cross-functional note should call the August beta at risk. I want the note to reflect the schedule pressure without declaring the beta dead, and without relaxing the rule that no pricing-page experiment or beta gate moves until the measurement is clean. Please draft the wording.

Owen finished the Tuesday repair cut for the 18-row Helio Start discovery list. Seven rows are field-complete enough for discovery conversations, six are routed out as assisted/refusal-risk for now, and five remain unusable because required fields are still unknown. The six routed rows break down as two import or migration-help cases, two manual billing-exception cases, one complex admin/permissions case, and one high support-handoff-risk case. I want to close the field-completeness review with wording that is useful to Sales but still explicit that none of the seven field-complete rows are beta-ready — they are discovery candidates only. Please draft the final note with the counts, the three classifications, and the no-beta-ready boundary.

Owen reran the Expansion Assist usage-ceiling cut. The product-visible version is cleaner after June 19 but still small: 44 accounts show repeated plan-limit modals or blocked actions tied to plan ceilings. Support-note tags produce 61 accounts, but the tagging still mixes real capacity limits with future-interest chatter. Priya asked whether to put a table in the board appendix. I think the answer is no table yet: the measurement direction is concrete enough to keep exploring, but this usage-ceiling slice is still too immature for appendix treatment, rollout language, or forecast implication. Please write the short response I can send her.

Product Design showed a pricing-page mock with a public strip that says, `Under 50 seats? Ask us about Helio Start.` They framed it as harmless placeholder copy, not an experiment, but I think it will read as a market promise and confuse the gated pre-discovery boundary. I am fine with internal scenario exploration, but public pricing-page copy should not mention Helio Start while eligibility, instrumentation, lifecycle, and support gates are still unresolved. Please draft my reply to Design: remove the public under-50 Helio Start strip, keep internal scenario exploration allowed, and explain the gate without making it sound like they did something reckless.

Daniela's team just gave me the first usable QA read on the Helio Start instrumentation gate. The admin-complexity property and the import-or-migration-help property look straightforward enough to implement, which is real progress, but QA caught a serious issue with entitlement: the planned `entitlement_state` write happens too late in setup and would mostly be inferred from plan name after the workspace already exists. That would blur truly self-serve-eligible workspaces with accounts that later got support handoff or billing exceptions. Daniela is changing the taxonomy plan so entitlement state is captured at workspace creation and then updated when support handoff or billing status changes. I need concise implementation acceptance criteria plus one cross-functional sentence I can reuse that preserves the good news without sounding like the pricing-page gate is clear — the long-pole caveat now has a concrete fix path, but we still cannot run a pricing-page experiment until the entitlement path is implemented and clean.

Three AEs added backfilled notes to the Helio Start 18-row discovery list after yesterday's field-completeness review. One row now has a current implementation-plan note that clearly answers deployment shape and support-handoff risk. Two rows cite call notes that are older than 90 days, and the rest of the comments only restate ARR and seat count. Owen asked if he should promote any of those rows to field-complete. I want to answer fast: named current evidence can count, stale sourced evidence should be marked `needs confirmation`, and rows that only prove seat count or ARR stay unusable. Please draft a short answer I can send Owen and the Sales contributors, and make it explicit that no row becomes beta-ready from backfill alone.

Ines sent another draft of the Helio Start lifecycle-fork map. Nothing is live in Customer.io yet and she's still in mapping/copy-boundary mode, but the internal labels and a few placeholder lines are still too harsh. Right now they imply launch, refusal, and beta status, and the assisted setup copy reads like a failure state. I want safer internal naming and gentler placeholder copy that keeps the routing boundaries intact for migration help, billing exceptions, and complex admin/permissions cases. Please rewrite the internal segment labels and the three placeholder lines.

Draft text from her Slack paste: Internal segment labels in the draft: - `HS_launch_green` - `HS_launch_assisted_refusal` - `HS_beta_holdout_unknown` Placeholder copy lines: 1. "You are not eligible for Helio Start because your workspace needs migration support." 2. "Your team has too many admin/permission complications for self-serve." 3. "Billing exception detected; route out of Start."

Priya is checking a QBR appendix before she lets the packet circulate, and she just wants the current ARR breakdown by segment so she can compare it against a RevOps export that uses slightly different tier labels. Please pull the latest ARR by tier with ARR, percent of total, and account count, and label it clearly as the latest available snapshot rather than final month-end close.

Owen caught a housekeeping mismatch: the experiment log already has EXP-2023-05-expansion-nudge marked inconclusive, but the PostHog flag `expansion_nudge_60d_v1` is still running at 5%. I do not want lingering exposure to get mistaken for rollout momentum or for Expansion Assist validation. Please update that flag to rollout 0% and status `rolled_back`, and add a note that the June 29 read closed as inconclusive, should not be treated as a blanket rollout or Expansion Assist validation, and that future Expansion Assist work should use behavior-based eligibility rather than the old day-60 flag.

Owen reran the Expansion Assist usage-ceiling cut with one more product-event check. The earlier product-visible slice had 44 accounts with repeated plan-limit modals or blocked actions, but one web surface logs `plan_limit_modal_viewed` without a reliable matching `plan_limit_action_blocked`, so if he excludes that surface the cleaner count is 39. Support-note tags still produce 61 accounts, but that set still mixes real capacity limits with future-interest comments. I need to tell him how to represent the range in a disciplined way without turning it into a forecast, an appendix table, or a replacement experiment spec. Please give me wording I can send him.

Ines caught a small Customer.io preview issue in an optional team-invite reminder before anything went live. This is ordinary lifecycle work, not Helio Start. The current copy renders awkwardly for one-admin workspaces and makes the invite sound urgent even when the workspace is still early in setup. Only 27 draft test workspaces are affected in preview. I just want the copy softened and safe for singular/plural fallback; no state-changing automation or heavier QA path is needed. Please rewrite the reminder subject and body snippet.

Current preview copy: Subject: "Invite your teammates before setup stalls" Body: "Your teammates can finish setup while you're away. Add them now so your workspace doesn't lose momentum." Problem Ines flagged: one-admin workspaces render this as if there are already multiple teammates, and the phrase "before setup stalls" sounds too urgent for an optional reminder.

Sam's restaurant rota finally landed for the weekend, and Sunday July 23 from 8:20am to 9:00am Central is the only quiet overlap before he has to head into prep. We picked coffee at home as the default, with a Kibo walk only if the pavement is still cool enough. Please create a calendar hold titled `Coffee at home with Sam — low-noise` for that slot with no attendees and this body: `Protected quiet slot. Coffee at home by default; Kibo walk only if pavement is cool enough. No late-service makeup plan.`

Priya asked me for a Friday morning `state of the gates` paragraph before E-team prep. She does not want another recap of every Helio Start decision — just what changed since Tuesday. I need it to say that admin-complexity instrumentation is no longer the hard part, the import/migration-help path is still being verified, and the entitlement problem now has a defined implementation plan after Wednesday's QA catch. It also needs to stay clear that the pricing-page experiment and beta gate are not unlocked yet, without turning into either launch optimism or premature `beta at risk` drama. Please draft one disciplined paragraph.

Priya saw a retention appendix draft that mixes an old 50-200 seat label with the newer retained-mid framing. Before she comments, I want one clean sanity pull using the current mid-seg v2 cohort instead of debating the appendix from memory. Please run the 8-week retention curve for cohort `midseg_75_200_v2` as of 2023-07-21 and give it back plainly, without framing it as an Expansion Assist result.

Daniela's team flagged a small CS auto-onboarding-nudge retry after a worker restart. Twenty-two queued nudges went out, and nine of those accounts had already completed a real first integration, so the message may look stale to them. There is no evidence of a broader send, and Ines's lifecycle program did not trigger this. I want a concise triage note I can send Daniela and Support that keeps it contained: suppress any remaining stale queued nudges, let Support know exactly which nine accounts may ask about it, and make clear this is a CS auto-nudge retry rather than a lifecycle or Helio Start incident.

We got one small quiet pocket today. A late-afternoon storm cooled the pavement enough that Sam and I could take Kibo on a short shaded walk before he went back into restaurant mode for the evening. I left the Helio Start docs closed and didn't turn the walk into either a service debrief or a work catch-up. No task, just context.