What it is in one line
A private Mac app for managers — not for meeting notes, but for the judgment you hold about
each person you lead: write down what you believe (with a confidence level and a status), tag
the signals you notice in your 1:1s, let those signals revise your beliefs on the record, and
watch stale signals fade with time.
Who built it
Chris Messina — well-known product person who coined and popularized the hashtag, later at
Google and Uber. He says he managed engineering teams for years and kept hitting the same
problem: he'd notice something about a direct report (a dip in engagement, a change in how they
handled ambiguity, a theory about the cause) and then lose it. Not the meeting notes — the
reason underneath, which drifted until a review forced him to rebuild it from the last few
weeks. He tried Notion, Apple Notes, and a spreadsheet; none stuck.
Read: a product born from being tortured by your own memory. And the pattern fits Messina's
playbook — the hashtag and his related products were mostly "give the thing that used to live
in people's heads a container." Continuum is the management version of that same move.
What it actually does
- Belief management → for each person, write the judgment you hold, with a confidence
(uncertain, likely, plausible, certain) and a status (active, watching, resolved, discarded)
- Signal tagging in 1:1s → when writing up a 1:1, tag the signals in it (productivity,
ownership, proactiveness, etc.), each with a direction: up, down, or steady
- Signals feed beliefs → signals flow into the beliefs they touch, nudging confidence and
leaving a note about what changed and why
- Time decay → a signal fades unless you observe it again, so what surfaces is the present
read, not a three-month-old snapshot
- Deliberately absent → no scores, no performance reviews, nothing reported anywhere; the
data never leaves that Mac
What it deliberately does not do: it is not an HR tool — no 360 reviews, no generated reports
to anyone.
What old behavior it replaces
A manager's read on each report used to live in the head: observations from meetings, theories
about motivation, who is shifting in which direction — scattered across memory and loose notes,
rebuilt painfully from the last few weeks whenever a review came around. Notion, Apple Notes,
and spreadsheets were tried, but they are general tools; "track a judgment as it evolves" had
no dedicated form.
Continuum replaces the drifting in-head judgment about a person — it makes the hypothesis
explicit, keeps it updated by signals, and lets old evidence decay naturally. It is not note-
taking; it is version control for judgment.
Business model
Freemium: free up to 3 people; Pro at $12/month or $108/year (14-day trial on yearly), or a
one-time $249 lifetime. Launch-week discounts: 50% off the first year monthly, 50% off yearly
with code CONTINUUMPH2026, 50% off lifetime.
Read: a free tier capped at 3 people is a smart threshold — a manager usually carries 5-10
reports, so natural use hits the paywall within days. The $249 buyout settles the "it's just a
small tool" objection in one shot. But it is fundamentally a single-user personal product, and
the unvalidated assumption is whether managers will pay separately for managing-people memory.
Hard numbers
- Launched at launch 2026-08-11: 131 upvotes / 8 comments, #7 on the daily leaderboard
- Launch topics: Mac, Productivity, Meetings
- Paying users, retention, revenue: not disclosed
- Team: not disclosed (product shape implies a small solo build)
- Pricing: free (≤3 people) / $12 per month / $108 per year / $249 lifetime
Four-way read
| Dimension |
Call |
| Founder-product fit |
Years of team management, real pain, tried every general tool — fit is maximal |
| Product insight |
The belief + confidence + signal + decay model is genuinely thought through; "no scores" is a deliberate anti-HR-tool design |
| Execution quality |
Local-first v1 Mac app, restrained and private; but single-user only |
| Timing |
Managerial memory-keeping is a real but scattered need; local privacy is both the selling point and the growth ceiling |
The call
The product model is worth dissecting; the commercial case is unproven.
The belief-and-signal mechanism is the most valuable part: store hypotheses, not facts; let
evidence update them continuously; let stale evidence decay — which is exactly how human
judgment about people actually runs. Ported to client relationships, channel management, or
sales follow-ups, it becomes a usable relationship-memory system.
Three problems. One, willingness to pay: whether managers pay $12/month for "memory about
the people I manage" has never been validated. Two, single-user ceiling: the judgment lives in
one head and dies on job change — the real product is a team version, which doesn't exist.
Three, missing AI: manually tagging signals in every 1:1 is the biggest operating cost, and
without auto-tagging few people will keep using it — and the founder himself asks in the launch
post "what signal types would you set up," which says that loop isn't thought through.
What to watch next
① Conversion rate from free (3 people) to Pro, and 30-day retention — is the pain real enough
to pay for
② Whether a team / multi-manager shared version appears — the decisive bottleneck for a
single-user tool
③ Whether auto-extraction of signals from 1:1 notes gets added — without cutting the manual
cost, there is no long-term usage
What you can take from it
Product logic: any relationship-type information that lives in the head and drifts (clients,
channels, reports, candidates) can take the four-piece "belief + confidence + signal + decay"
model — it gives judgment a version history instead of a pile of notes. This model transfers
directly into a CRM to replace dead data like static customer tags.
Pricing structure: the free-until-3-people wall is copyable — cap the free tier at the
typical scale of one managed unit, and let natural use push you into paying, instead of
feature-crippling.
Verdict
Unproven (the mechanism is worth watching). Credible founder, real pain, clear thinking —
but a 131-upvote launch, a single-user Mac app, and an unvalidated willingness to pay don't
support a higher call. Its real value is the belief-signal-decay relationship-memory model;
copying it costs less than buying it.