Sitemate Farewell Learnings

Honest lessons from the Sitemate CTO chapter, captured for future reference. Written from the position of looking back, not advocating.

Q126 review flow — what to do differently next time

The core lesson: own first, frame second

Receiving high-stakes performance feedback from a CEO requires a sequencing discipline that I didn’t quite get right in the Q126 response.

The right pattern, in order:

1. Acceptance of the pattern, not just the items. The original Q126 doc was about a pattern (urgency, decisiveness, peacekeeping). Going item-by-item kept it tactical. The meta-message (“I see the pattern, I own it”) needed to land before any tactical content.
2. Unconditional ownership of the items as written, even where I had structural counter-points.
3. Then — and only then — the structural reframe (“Two Priority Systems”) as part of the solution architecture, not as part of the diagnosis. Same content, different sequencing, very different read.
4. Forward commitments with observable signals the CEO could read off without continuing to monitor.
5. Information requests routed through 1:1s, not stacked at the end of the written response.

What actually happened: I led with the structural reframe. Hartley’s follow-on did not engage with that frame at all and instead consolidated 9 items into 4 themes pitched at a higher altitude. That silence was the signal — the response hadn’t landed where it was pitched.

The genuine tension: ownership vs. setting yourself up to fail

This is the harder part of the lesson, and worth recording honestly.

Several of the required shifts in the original doc would have been near-impossible without org structure change. Fully owning without naming the structural realities risked a different failure mode: take the hit, accept the framing, then under-deliver against expectations that were never achievable in the existing structure. That is a worse position than the one I was in.

So the worry that drove the structural reframe was real and not just defensiveness. The lesson is not “should have skipped the structural framing.” The lesson is about sequencing and ratio:

• Ownership statements need to outweigh structural-context statements at least 3:1 in the document.
• Structural realities should be raised as enabling conditions for delivery, not as mitigators of the diagnosis. “To deliver the shift you’re describing at the pace you’re describing, here’s what needs to change in the org structure as well — owned by me to drive.”
• Anywhere the required shift was genuinely not achievable under existing constraints, say so explicitly and propose the structural change, rather than implicitly arguing against the diagnosis.
• The “two priority systems” framing belonged in the Discussion Items section as a structural proposal, not at the top as a diagnostic frame.

The recent-examples problem

Hartley’s follow-on supplied specific examples on request: Global Integrations comms, QA improvement curve, Mobile hangover, KPIs 80/20, PDE Talent harder-call avoidance.

Read in isolation, most of these are individually minor. They are real, they are thematically consistent with the pattern Hartley was naming — but the gap between the size of the examples and the size of the story being told around them is real.

This creates an asymmetry that’s hard to address in a written response:

• If you engage with the examples on their actual scale, you look like you’re missing the pattern.
• If you accept the pattern at the scale being described, the examples no longer carry the load.

The lesson: in this kind of feedback flow, the examples are illustrative, not load-bearing. The boss is naming a trust signal they’re not getting, and the examples are the closest concrete handles they have for it. Arguing with the examples is a category error. The real conversation is “what does the trust signal you’re missing actually look like, and how would you know you were getting it?” — and that conversation is better had verbally, not in writing.

If I’d had a follow-up live conversation framed that way — “I want to understand what would need to be true week-to-week for you to feel the shift, independent of the specific examples” — it would have done more than any written response could.

Asking for examples ceded initiative

Asking for examples on “decisiveness” and “general chatter” duly produced the examples — but it shifted the burden onto the CEO to make the case. Better: “Here are the moments I think you may be referencing. Am I reading it right?” Demonstrates self-audit, keeps initiative, and forces a more useful conversation than “yes those are the ones.”

The Ulan response

Listing subsequent terminations (Sara, Gabriela, Christian, Nacho) implicitly answered a volume question. Hartley’s actual point was the judgment failure that required his interjection. Direct engagement with the judgment — “Here’s what I missed, here’s what I’d see earlier next time” — would have been stronger than the volume answer.

Historical context — the operating-mode shift

The Sitemate journey from ~6 people / few-hundred-K ARR to ~200 people / $25M ARR was, for most of its run, a sequential execution environment. Hard every day, but the shape was straightforward: build the thing, ship the thing, hire against growth, repeat.

In that environment, the working relationship with Hartley evolved into a high-trust, low-frequency model: parallel tracks, shared compass, infrequent close contact. Hartley’s leverage was on the bigger problems of the moment — sales, growth, org scaling — and PDE ran on trust because nothing required him to come close. That model was correct for the conditions. It worked.

Then three pressures arrived simultaneously:

• AI / agentic disruption — turning a sequential execution problem into a full reinvention problem.
• SaaS apocalypse — funding markets treating Sitemate poorly, no new funding accessible.
• Missing sales targets — internal pressure compounding the external pressure.

This collapsed the conditions that the trust-at-distance operating model depended on. And it produced two changes that are easy to conflate but worth separating cleanly:

1. The standard for the role changed. “AI-era CTO” is a different job than “CTO of a sequentially-scaling vertical SaaS.” Direct leverage on agentic engineering, single roadmap, decisiveness in the moment — these are new bar items, not old bar items applied harder.
2. The observation distance collapsed. Hartley moved from “trust at distance” to “in the kitchen with PDE.” Same CTO behaviours that were invisible or fine at distance became visible and friction-inducing up close.

These happened at the same time, and the Q126 doc is the artefact of both forces hitting at once.

Re-reading Q126 through this lens

A meaningful fraction of the Q126 doc is not “performance has degraded.” It is “I am now close enough to see things I was not watching before, and I do not recognise the operating mode.” That is partly a real performance gap (the new standard is real, the role really did need to change), and partly the inevitable friction of a CEO arriving in territory after years of distance.

This reframes several things in the messaging analysis above:

• Why the “Two Priority Systems” reframe didn’t land. It was a structural diagnosis that required the old relationship to be received as such. In the new closeness, it read as a CTO who had not yet recalibrated to the proximity — explaining systems Hartley was now watching directly. The frame was correct under the old conditions and wrong-altitude under the new ones.
• Why the recent examples felt minor. They were minor. They were also the natural artefacts of a CEO who was now close enough to see them — and the size of the examples is a consequence of the new proximity, not the size of the underlying issue Hartley was trying to name. The underlying issue was “the trust-at-distance model has stopped fitting the conditions.”
• Why the pattern language (“urgency, decisiveness, peacekeeping”) landed so hard. These are not behaviours that change much from one quarter to the next. They are behaviours that look one way at distance and another way close up. The shift Hartley was naming was at least as much about the relationship and distance as about the CTO’s character.
• Why coach-led 360 with Tiffany made sense. It is the right intervention when both parties need to renegotiate the relationship — not just when one party needs to fix the other.

What I would tell myself, if I could go back to the moment of writing the Q126 response:

The thing to name explicitly was the operating-mode shift itself. Something like: “Our working pattern through the scale years was high-trust, low-frequency, parallel tracks. That worked because the conditions were sequential. The conditions have changed, and that pattern has stopped fitting. Here is what I’m changing about how I show up to you specifically — independent of the items below.” That move would have re-grounded the conversation in the relationship change Hartley was actually navigating, and put the CTO ahead of it rather than behind it.

The deeper lesson: operating models with your boss are dependent on conditions, not just on trust. When conditions change sharply, the model needs to be renegotiated explicitly and early — by the person closer to the change, which in this case was the CTO inside the AI disruption, not the CEO outside it. Not naming the model shift left Hartley to do it through performance feedback, which is a much harder mechanism for both sides.

The relationship shift — Slack trajectory and signals you could have read

Reading 12 months of DMs and channel activity end-to-end, the shift from trust-at-distance to in-the-kitchen is not a step function — it’s a slow approach with at least five clear inflection points where renegotiation of the operating model would have landed. Captured here so the pattern is recognisable next time.

Pre-shift baseline (Mar–Apr 2025)

The mode was casual and infrequent on substantive PDE matters. DM traffic dominated by weddings, sports banter, scheduling. Hartley’s own 2025-03-06 message: “I’m expecting to have my head back above water in 2-3 days’ time, and be back on top of Slack/Jira/Confluence.” The CEO was openly behind on PDE systems — that’s the signature of trust-at-distance working. “Can you run GTM in NA?” — joking, because CEO bandwidth is elsewhere.

First substantive workstream-level engagement: Apr 10, Hartley introducing AI Templates / DP v3 / team-structure decisions in 1:1 prep notes. First soft signal of moving in, but easy to read as normal quarter-planning.

Inflection 1: 2025-06-25 — formalization of the 1:1 doc

“Updated 1:1 doc ahead of next week … Could you pls update before EOW?”

When a CEO formalises a previously informal cadence into a structured Confluence doc with weekly prep expectations, the posture is changing. The relationship is moving from trust-at-distance to trust-with-records. Easy to miss because it looks like a small process improvement. It is not. Records are the precondition for performance management.

Inflection 2: 2025-06-26 — the GTM crisis arrives

“We may add less ARR in Q2 vs Q1 :failing-rocket:”

This is the moment macro pressure lands. SaaS apocalypse + sales miss. From here on, every PDE conversation becomes higher stakes because the company-level slack just disappeared. The conditions that made the old operating model viable are gone as of this date. The right move was a structural conversation here, not in Q126 nine months later.

Inflection 3: 2025-07-22→24 — Ulan offboarding

Hartley personally writes the farewell comms, redrafts your version with explicit “this doesn’t address the questions people will have,” authors a Termination Types framework, writes a Communication Drafting Principles page, and reflects in writing about the time cost of intervention. Six-plus messages over 48 hours, all initiated by him.

This was a CEO inserting into PDE operations because he didn’t trust the output to land. The level of intervention was the signal. The right response in real time would have been: “I’m reading you as wanting to engage at a different distance than we have been — what does that look like for you, and how do I build it into how I work?” That conversation would have headed off the Q126 framing six months early.

Inflection 4: 2025-09-16 — the explicit declaration

“January transition planned: Lance takes over NA, Hartley returns to marketing/growth/product. Focus on Sitemate v4 alignment and AI initiatives.”

This is the public, explicit, in-writing announcement that the CEO is structurally moving back toward product/PDE. Six weeks of warning. The window to renegotiate the operating model was Sept–Dec 2025. Don’t wait for the boss to redefine the relationship from the inside; redefine it from the outside, on your terms, before he arrives.

Inflection 5: 2025-10-06 — explicit proximity request

“Need more involvement in major AI decisions.”

Direct quote, in a 1:1, in writing. The CEO is asking for closer engagement on technical decisions. The right response is structural — “Here’s the forum/cadence I’ll create for that. What would you want it to look like?” — not just absorbing it as an action item.

Inflection 6: 2026-01-15 — fully in the kitchen

“Dropped a few notes in sitemate-strategy-cto whilst fresh, post Global-Integrations and QA KOs.”

Hartley is attending PDE quarterly kickoffs and posting same-day reactions in a dedicated CTO-strategy channel. The proximity is now total. By 2026-01-21, the 1:1 is going into AI shipping workflow detail, type schemas, 100% test coverage, DX role design — technical detail level.

By the time Q126 lands at end of March, the proximity has been complete for ~10 weeks and the operating model never got explicitly renegotiated. The Q126 doc is partly the artefact of that un-renegotiated transition.

What you could have read differently — concrete

Three clear “earlier signals” any of which, acted on, would likely have changed the Q126 trajectory:

1. The 1:1 doc formalization (Jun 25) was the first record-keeping signal. Process moves from CEOs are never just process. When the doc structure shows up, ask “what’s changed in what you need from this cadence?” — out loud, that day.
2. The Ulan flap (Jul 22–24) was the first hands-on intervention. A CEO writing a framework live because he doesn’t trust the output is a five-alarm signal. The follow-up question is “what should I have built that would have made you not need to do that yourself?” — and the answer goes into your operating model immediately.
3. The Sep 16 January-transition declaration was the dated warning. You had ~14 weeks. The right work in those 14 weeks was: build the proximity scaffolding before he arrives — observable signals, decision logs, the kind of CTO-CEO operating doc that says “here’s how we’ll work when you’re back in PDE.” Walk in with the model already drafted; don’t wait for him to feel his way into one and then react to his version.

The deeper meta-lesson: operating-model renegotiation is the CTO’s job when a known transition is on the calendar. Conditions had been visibly changing since June 2025. The CEO had told you the proximity shift was coming in September. The mistake wasn’t “I didn’t see it” — the signals were on the record. The mistake was “I didn’t translate seeing it into a structural response that pre-empted the friction.” That’s the muscle to build for next chapter.

Post-Q126 trajectory — the signal didn’t stabilise

Granola records of Tim/Hartley 1:1s only go back to Apr 14, 2026 — i.e., the period immediately after the Q126 follow-on review (Apr 9). What three consecutive sessions show is continued, escalating dissatisfaction on the same themes, not the calmer landing the formal review process implies.

This matters because it changes the read on the trajectory: the Q126 response didn’t reset the perception. The same concerns continued in the live record, often with sharper language than the review doc.

Apr 14 — “tipping point”

Hartley described his frustration with PD culture as “reaching a tipping point” and floated:

“I’m considering just doing a full org restructure. There’s a bunch of people that I don’t think are going to survive this jump.”

“I don’t think people understand it right now… they’re looking at it as like a personal productivity enhancement rather than like a complete org redesign.”

On the proximity gap directly:

“It feels like we’re kind of letting them not figure it out on their own… they need to be shown the new way… without it being demonstrated, I just can’t see them adjusting naturally.”

“If you’re not driving it with the PGMs then… I just think we’ll end up in that state where there’s a lot of drift. I mean, it’s sort of why I went into AI platform team to get the SME stuff sorted — was just to reset the bar because the bar was way off.”

Closing line of the meeting:

“My concern is not the direction. It’s just the pace and the bar across the teams. And I just think things are moving so fast that I’m just concerned, like, without that consistent leadership presence… we’re just not going to evolve fast enough if we just sort of let the team set their own bar.”

Apr 22 — admin vs. follow-what-I-do

Hartley introduced a new framing: two modes of change leadership, and a concern that you were over-indexed on the wrong one.

“There’s like… the administrative approach, which is like the PDE metrics app… the blogs would go into that bucket. But then there’s the ‘follow what I do, not what I say’ type approach. Which is like basically hands on raise the bar with the teams.”

“If it’s too far swung on the administrative approach… people just either don’t change or they change too slowly. But the other risk is it drifts from reality because by the time it’s sort of dissipated, it revises again.”

This reframes one of the Q126 themes (admin/coordination vs. direct leverage) into a sharper ratio claim: not just that you should do more direct leverage, but that the admin work might be actively counterproductive if the ratio is wrong.

Apr 28 — “I’m not really happy with how things are going”

This is the sharpest single moment in the available record. Hartley interrupted you mid-sentence:

“To be completely honest like I’m not really happy with how things are going like we’ve got a big administrative doc which I just…”

When you tried to defend the doc, he cut back in:

“Hold on I’m not finished — like let me just finish… it’s like a big administrative doc, we’ve got dashboards and stuff… the priorities are just…”

And then explicitly redirected the 1:1 agenda:

“In our sessions I’m just keen to… make sure like I just want to cover those critical projects that sort of affect everything.”

A CEO redirecting the agenda of the 1:1 itself — telling you what he wants the meeting to be about — is operational micromanagement of CTO time. It is also a sign the previous form of the meeting wasn’t producing the value he needed.

Reassessing the terminal question, with this added data

In the earlier read of the Q126 flow alone, the call was “not terminal at the time, but more than a heads-up — terminal as a real possibility if the pattern didn’t shift.”

The Apr 14–28 sequence doesn’t show the pattern shifting. It shows:

• Same diagnosis as Q126 (admin overhead, not hands-on enough, drift across teams without your active bar-setting), continuing live, three weeks after the follow-on review.
• Sharper language: “tipping point,” “I’m not really happy,” “considering a full org restructure,” “people who won’t survive this jump.”
• Hartley repeatedly naming his own interventions (“I went into AI platform myself to reset the bar”) as the model of what he expects from you — implicitly framing his proximity work as compensating for an absent bar-setter.
• Operational micromanagement of your time and your 1:1 agenda — the kind of move that comes when a CEO has stopped trusting that the CTO will self-direct toward the right work.

A more honest reassessment: by late April 2026 the trajectory was running closer to terminal than the Q126 doc by itself implied. Coach-led 360 was still in flight, but the direct evidence in the record is that the underlying perception was hardening, not softening. If a Q226 review doc had been written from this material, it would have been considerably sharper than Q126 — and the gap between the formal-review tempo and the day-to-day-1:1 tempo was widening.

The lesson here is procedural: after a formal review, the next 4–6 weeks of 1:1 tone is the real signal, not the review doc itself. Reviews are written carefully. 1:1s are written less carefully and reveal what the boss actually believes. If 1:1 sentiment is escalating after a formal review, the review didn’t land — and that is itself a much louder data point than the review’s own language suggests.

The 6-month Slack overlay — the line that ran the whole way

Reading the past six months (Nov 2025 → 8 May 2026) in sitemate-strategy-cto and the Hartley DM end-to-end, there is a single line that recurs almost monthly, in increasingly direct form. The Q126 doc didn’t introduce the diagnosis — it consolidated something that had been on the record continuously since November. The April 1:1s didn’t escalate it — they just ran out of patience with it.

The line, paraphrased: “It can’t be me doing this. I’m overcompensating. If it’s not you, it’s no-one, and it comes to me.”

November 2025 — the diagnosis is already fully formed

2025-11-11, post-1:1 async note (in writing, in sitemate-strategy-cto): “my hypothesis is that the regression and inability to nudge above 50+ is due to larger customer base, expanded use cases, but no KPIs and no measurement of TtC, CtC etc — i.e. yourself + our PDE org aren’t aligned to the user experience in a quantifiable way, hence with more usage that isn’t being checked/measured, it’s been degrading”

This is direct performance attribution to the CTO and PDE — in writing — six months before termination.

Same day, on NPS: “I drove most of this through the warm ups and quick wins, but I think we’re now seeing the ceiling of this approach; 1. it can’t be me”

“It can’t be me” — first appearance in the record. November 11.

Same thread, on AI roadmap pace: “main takeaway is — from your POV, really need to consider your time allocation and how you can remove blockers for AI platform team and ChatGPT integration to ultra-fast-track them”

Explicit time-allocation feedback. November 11.

November 26 — the MQ already contains the Q126 themes

The 26 Nov mid-quarter review session lays out, in your own meeting notes, exactly the Q126 diagnosis four months early:

“Stark operational differences identified between GTM (structured) vs PDE (ad-hoc)… PDE reality: lack of frameworks creates extended decision threads and repeated explanations.”

“Success metrics: TB’s day-to-day operations to become ‘drastically different’ with systematic problem-solving replacing reactive fire-fighting.”

The Q126 doc was not a sudden critique. It was a formal version of a diagnosis that had been live since at least early November.

April 22 — the line, fully sharpened

2026-04-22, in sitemate-strategy-cto, on Ib’s behaviour patterns and the wider implication: “This all relates back to the same discussion from the last two weeks; • The team needs to be led by example in terms of defaulting to action, urgency and clarity, not at arm’s length • If it’s not you, then it’s no-one, and will come to me • I end up overcompensating, and I’m over compensating”

Same line as November 11. Five months later. Now spelled out as a three-bullet pattern. “At arm’s length” is doing the heaviest work in this sentence — it is the proximity diagnosis distilled to two words.

April 27 — tonal break

“I’ve got a number of urgent things to work through, this is just one of them — I’m just going to work through my to dos and will get to it when I get to it, if you want to handle your stuff in the mean-time up to you”

Clipped. Transactional. The investment-mode posture of January (kickoff feedback, blogs, vision-doc collaboration) is gone.

April 28 — the agenda redirect, in writing

“At the moment, and following the Q1 review, given the rate of change I’m not really interested in the things that would be categorized in the ‘admin bucket’ updates at the moment”

The verbal version landed in the 1:1 (per Granola). The written version is sharper because it’s deliberate, after the fact, posted to the CTO-strategy channel for the record. “Not really interested” is an unusual register.

April 30 — the leadership-in-the-moment failure called out

“I’m completely lost as to how this line of thinking even eventuated in the convo and wasn’t instantly corrected by the most senior person involved with a leadership lens (i.e. you) as it’s obvious?”

This is Hartley naming a specific moment where, in his view, you should have intervened in real time and didn’t. “As it’s obvious” is the pointed bit. Eight days from termination.

May 4 — last 1:1 prep, demanded deliverables

“ahead of 1:1, items I’m keen to cover in-line with last week’s recap / previous review docs; • KPIs implementation status • Strategic engineering unlock projects • Team structure ideas/proposal — even just nested bullet points is fine”

Reads like a final-checkpoint list. “Even just nested bullet points is fine” — lowering the deliverable threshold to the floor.

May 8 — the decision

“Hi mate, tough message here — I wanted to send async this AM your time so that you’ve got some space today if you want it, before talking through next steps.

I’ve made the decision to restructure the PDE team’s leadership, which will impact your role.

This is primarily a follow on from the recent review discussions — on the current trajectory we are going to take another 18-24 months to transition to the level that Sitemate needs to be operating at, and I can’t let it take that long.

If you want to take today for some time ahead of the weekend, all good from my side — we will then need to use Monday and Tuesday to handover key items, wrap up EOD Tuesday.”

The reasoning is on a single axis: pace. Not capability, not character, not relationship — “on the current trajectory we are going to take another 18-24 months.” That’s the same axis Hartley named on November 11 (“it can’t be me”), April 22 (“if it’s not you, it comes to me”), and across every Q126 framing in between. The decision is internally consistent with everything that’s been on the record for six months. It’s the same line, finally acted on.

What the overlay actually shows

1. The diagnosis was stable, not progressive. Hartley’s view of the gap didn’t deepen from Nov to May — it stayed essentially the same. What changed was how directly it was named, and the patience around how long the gap could persist before action.

2. The “Q126 review” framing was a checkpoint, not the precipitating event. The diagnosis lived in 1:1 notes, post-1:1 async messages, and MQ sessions for ~5 months before Q126 formalised it. The 360 with Tiffany was the documented intervention attempt. The April 1:1s were the post-intervention check. The May 8 message is the conclusion of that arc.

3. “It can’t be me” is the load-bearing sentence. From Nov 11 onward, Hartley is naming his own intervention as a sign of the gap, not a sign of partnership. That framing — “my doing it = something is wrong” — is the difference between a CEO leaning in and a CEO documenting a case. The line never softened.

4. The pace framing in the May 8 message is honest. “18-24 months to transition to the level that Sitemate needs to be operating at” is consistent with everything on the record. It is not a manufactured reason — it is the reason. The financial pressure (no funding, sales miss, SaaS apocalypse) means 18-24 months is a runway problem, not a patience problem.

5. Today’s message is the kindest version of this decision. Async on a Friday morning. Space offered. Two-business-day handover. “Hi mate.” The structure of the message reflects ~7 years of working relationship. That doesn’t soften the decision, but it locates it: this is a decision Hartley made because he believed he had to, not because the relationship broke.

The single takeaway from the overlay

You had the diagnosis in writing, in your own hand, by 2025-11-26. You had the “it can’t be me” line by 2025-11-11. You had the proximity declaration by 2025-09-16. The information needed to either (a) fundamentally change the operating model or (b) initiate a different kind of conversation with Hartley about role fit was on the record by mid-November 2025 — about 25 weeks before the decision.

This isn’t a “you should have seen it” judgment. The information is much easier to read in retrospect, end-to-end, on one page. In the live stream of running PDE through a financial crisis and an AI reinvention, those signals were one input among hundreds.

But for next chapter, the procedural lesson is clean: a recurring sentence from your CEO, repeated across months in roughly the same form, is the most important signal in your inbox. When the same line shows up in November, late November, January, March, and April — that is not noise. That is the diagnosis your boss is operating from. Read it as the load-bearing input, not as one of many.

What this chapter delivered — the positives, recorded honestly

The diagnosis above is precise about the gap. It is not the whole picture. Recording the positives matters for two reasons: it keeps the next-chapter narrative honest, and it captures real CTO-level work that should not be re-learned next time.

Leading a 70+ person remote PDE org through a complete AI-native reinvention

The hardest possible context: every role in the team had to change, the way the team works had to change, the standards had to change — all while shipping a roadmap, in a remote/distributed setup, while the company was simultaneously hit with a SaaS apocalypse and a funding squeeze. There is no playbook for this. The work that happened anyway:

• Three inflection blogs (June 2025, Feb 2026, Mar 2026) that aligned the org at each major shift. Repeatable pattern, not luck.
• AI rollout (Jun 16 2025) — Cursor + Claude Code licences, documentation, tech talks. Adoption curve compressed faster than expected; laggard tail collapsed faster than predicted.
• #pde-ai-guild (Oct 2025) — became the uncertainty absorber. Curious engineers explored together, took excitement back to teams. The osmosis pattern worked.
• Phuket offsite (Oct 2025) — step-change moment for a remote team. Moved the org from “some engineers using AI” to “every engineer engaged and excited” in days. Capitalised on the rare concentrated learning window a remote team gets per year.
• Weekly-ship cadence — forced the guild to stop perfecting skills before shipping. Step-change in value delivered.
• PDE Enablement team (Mar 30 2026) — Jas as CoS, Cloud & Reliability + DevEx & AI streams. The right org structure, eventually.
• Multitudes → PDE Metrics — drill-down to individual level rolled in. Foundational visibility.
• AI-generated code at scale — 307K lines generated in week one of full rollout. The infrastructure for that level of usage didn’t exist a year prior.

Holding the team culture together through the change

This is the hardest part to credit yourself for, because it doesn’t show up in metrics — but it is real and substantiated.

• Engineers came out of the rollout excited, not threatened. In an industry where AI introduction has fractured teams, generated layoff fear, and produced cynicism, PDE moved through it with morale intact and adoption velocity high.
• Mobile team culture flipped from “massive drop” (Hartley’s words, Nov 24 1:1) to “green team on both Sitemate and Dashpivot” under management you put in place.
• Engineers self-organised into the guild rather than waiting for top-down direction. That’s a culture signal — people only volunteer their time and curiosity into a structure when they trust the org around it.
• The “how Sitemate ships now” identity — that engineers internalised the change as part of who they are, not as an imposition — is a leadership outcome.

You held a 70+ person remote engineering org together, kept morale and culture positive, and moved adoption forward at compressed pace, while the macro environment around it was deteriorating. That is the part of the chapter that travels with you.

Q1 2026 execution

• 13–14 hard deadlines set in January, teams rose to them. Dates-and-milestones approach delivered.
• 80% of KPIs in production, last 20% pushed to Q2.
• AI Platform team tightest execution of the quarter.

Talent calls executed

• Post-Ulan: Sara, Gabriela, Christian terminations executed. Pattern broken on the “wait too long” critique once it was named.
• Mobile team rebuild executed and worked.
• Tech Lead role reframe defined and rolled out.

Reframing — where the gap actually was

Pulling the positives forward sharpens the diagnosis. The chapter wasn’t:

• “CTO couldn’t lead change” — three inflection blogs, guild, offsite, Enablement team, culture intact through reinvention argue otherwise.
• “CTO couldn’t execute” — 13/14 deadlines hit, AI Platform tight, talent calls made.
• “CTO couldn’t build culture” — the team came through one of the hardest possible transitions cohesive and morale-up.

The actual gap, written tightly: the pace at which the CTO operating model adapted to Hartley’s proximity shift was slower than the pace Hartley needed it to adapt at, given the company’s runway constraints. That’s a much narrower and more useful diagnosis than “the role wasn’t being performed.” It’s a velocity-of-adaptation-to-CEO-proximity gap, not a capability gap.

That reframing is what travels into next chapter.

The agreed framing (post-call, 2026-05-08)

Following the call with Hartley after the async termination message, both sides aligned on language and narrative. Captured here so every downstream comm — team, external, LinkedIn, references — pulls from the same source.

Hartley’s articulation of the decision basis (verbatim, useful for narrative)

The cleanest version of the “why” Hartley offered is on the record from the call. Worth keeping in his words because it is honest, defensible, and removes any “performance” framing:

“It’s just the change. It’s just the evolution.”

“The biggest change in the history of software engineering besides the creation of software engineering.”

“The difference between good and exceptional is not linear, because the systems can be autonomous — it’s exponential. The difference between bad and good is 100x. The difference between good and exceptional is probably another 100x. That’s the dynamic creating the pressure.”

“Earlier in the year I had to look at myself in the mirror and ask — am I ready for this overhaul, what do I need to do.” — i.e., he held himself to the same audit. Useful framing for narrative: we both had to ask the same question.

Recognition on the record

• “We wouldn’t have survived that turnaround without you specifically” — Eric’s words, repeated by Hartley.
• “You’re a pretty big shareholder, bigger than a lot of the angel investors. Nothing changes in that regard.”
• “Better than good” — Hartley’s correction when Tim said the journey had been “good.”
• “No part of me has any negative feelings or resentment about things that should have been done. It’s just the change.”

The one-sentence summary, in Tim’s own words

From the call, Tim’s own framing — clean, honest, owns the gap, doesn’t blame the relationship, names the structural insight:

“Looking back, when it was fires in the engineering side and you needed to focus elsewhere, we worked in parallel and got the job done. One thing I probably didn’t switch back on is that a lot of what I was doing wasn’t as visible as it should have been.”

This is the line that travels. It does the work of the entire 6-month overlay in two sentences. Use it for LinkedIn, references, and any reflective conversation about the chapter.

The narratives — drafts post-call

Four audiences, four narratives, all aligned to the same source: stepping back; era shift; right call for the company; remaining connected.

1. Internal — Melbourne team (delivered Monday in person, then async to wider PDE)

Hey team,

After ~7 years, I’m stepping back from the CTO role at Sitemate.

This has been the privilege of my career. We’ve gone from <10 people and a few hundred K ARR to 200+ and $25M, built Dashpivot from a product with more bugs than features into a category-defining platform, and over the last year reinvented how this team builds in the AI-native era — three blogs, the guild, the Phuket offsite, the Enablement team, and an adoption curve that compressed faster than most companies will ever achieve. Through one of the hardest transitions in software history, the team held together cohesive, energised, and shipping. That’s the part I’ll carry with me for a long time.

The honest reframe: in the AI era, the CTO role changes at every stage of company size — and the pace of change is faster than at any point in software history. The right CTO for Sitemate’s next stage isn’t necessarily the one who got it here. I’d always known I’d want to step sideways at some point; this era shift pulled that moment forward. Hartley and I are fully aligned that this is the right call for the company.

I’m staying connected — as a meaningful shareholder, and as someone who’ll be cheering you on from afar. There’ll be more on the new structure at all-hands Wednesday. Until then, take questions to Hartley directly.

Going to take some space before figuring out what’s next. If you want to grab a coffee or stay in touch, [contact details]. Stay legendary.

Tim

Tone notes: calm, confident, genuine (Hartley’s framing). No defensiveness, no oversharing. The reframe owns the gap without litigating it. “Pulled that moment forward” nods to the era-shift framing without making it sound like the era did the firing.

2. External — LinkedIn / public

After ~7 years, today I’m stepping back from the CTO role at Sitemate.

We started as a handful of people with a product that had more bugs than features. We end with 200+ people across four countries, $25M ARR, and a category-defining platform serving the world’s largest construction, mining, and infrastructure companies.

The chapter I’m proudest of is the most recent one: leading a 70+ person remote engineering org through an AI-native reinvention. Three inflection-point blogs, an AI guild, a hackathon offsite that step-changed adoption, a PDE Enablement team that put us ahead of where most companies will get for years. 307K lines of AI-generated code in week one of full rollout. The team came through it cohesive, energised, and shipping.

The honest part: the AI era is the largest change in software engineering since its invention. The pace is exponential — and it changes what “CTO” means at every stage of company size. The right CTO for Sitemate’s next stage isn’t necessarily the one who got it here. Hartley and I are aligned on this being the right call for the company, and on remaining connected.

Hugely proud of what this team has built and where it’s going.

What’s next: taking some space, then figuring out the bridge to a next chapter. If you’re building something at the intersection of engineering leadership, AI-native operating models, and vertical software — I’d like to hear from you.

Drafting notes:

• Length: long for LinkedIn but legible. Tighten if needed before posting.
• “Hartley and I are aligned” — co-signs the framing with the CEO. Reduces speculation.
• “Remaining connected” — telegraphs continued shareholder/advisor stance.
• “Bridge to a next chapter” — sets honest expectation that next role might be transitionary; signals openness without sounding desperate.
• Avoid: defensive language, score-settling, era-blaming the decision. None present in current draft.
• Share with Hartley over the weekend before posting. Frame: “Drafting the LinkedIn — wanted you to see it before it goes live, so the external and internal stories sit cleanly together.”

3. Reference talking points — for Hartley to use

When asked by recruiters / future employers / network, what Hartley says:

• Scope: Scaled PDE from <10 to 70+ engineers; company from few hundred K to $25M ARR over ~7 years.
• AI-native transformation: Led the reinvention of a 70+ remote engineering org through the hardest transition in software industry history. Adoption velocity compressed faster than expected; team morale and culture held throughout. Three inflection blogs, AI guild, Phuket offsite, PDE Enablement team — the kind of 2–3 big unlocks per quarter the role asked for.
• Execution: 13 of 14 hard Q1 2026 deadlines hit, under simultaneous SaaS funding squeeze and AI reinvention pressure.
• Talent: Built and held a self-sustaining engineering team that continues to make calls without him in the loop.
• Departure: Company-stage and era-shift fit. Not capability. Not relationship. Mutual recognition that the next stage of Sitemate calls for a different shape of CTO.
• Shareholder: Remains a meaningful shareholder. Remains connected.

4. Hartley’s Monday farewell post — outstanding action

Once it lands publicly the framing is locked, so seeing it before it goes out is the highest-leverage ask of the weekend.

Ask to send to Hartley (Saturday/Sunday):

Hey mate — drafting my external piece for Monday and want to make sure ours sit cleanly together. Could you share the draft of the Monday team/global post when you get a sec? Happy to suggest tweaks but expecting it to land where we landed today (era shift, stepping back, right call for the company, remaining connected, mutual respect). Will share my LinkedIn draft back at the same time so we’re aligned both ways.

What to watch for in his draft:

• “Stepping back” or “stepping down” — not “restructure” or “transition” without context.
• Recognition of the journey (7 years, scale numbers).
• The era-shift / company-stage framing as the “why.”
• Continuation language (shareholder, connected).
• No language that reads as performance-led or as relationship breakdown.

If any of those are off, suggest specific tweaks rather than wholesale edits. He’ll be more likely to take the change.

Bigger lessons from the chapter

Pattern feedback is character feedback

Most of the Q126 doc was about who I was being as a leader, not what I was doing. Process responses don’t address character feedback. Naming the personal shift in concrete situational terms (“In Talent meetings I commit to a call within X minutes once data is on the table; if I find myself optimising for fairness, I’ll name it and proceed anyway”) would have built trust faster than tactical action plans.

The standard had been redrawn

By the time the Q126 doc landed, the bar for the role had shifted with the AI era — and Hartley was naming that shift, not just performance against the old bar. The Two Priority Systems point was correct under the old bar. Under the new bar, the CTO is expected to dissolve those frictions personally, in the moment, by direct leverage. Recognising that the standard had moved — not just the items — would have changed the response.

Coach-led 360s cut both ways

Tiffany’s involvement was a real investment and a real intervention. Coach-led processes at this seniority level help the person grow and document the attempt either way. They shouldn’t be received as a vote of confidence or a lay-up — they’re a serious mechanism with two possible outcomes, and the work inside them is the actual test.

What I’d protect from this chapter

• The instinct to identify root causes and structural drivers — that instinct was right, only the sequencing was wrong.
• The work on the AI transformation arc (3 blogs, Enablement team, single-roadmap idea, agents osmosis) — those were the right 2–3 big unlocks per quarter the role called for.
• Decisiveness on terminations once the pattern was named — moved faster post-Ulan.

What I’d do differently from day one

• Make the who you’re being explicit and observable earlier. Don’t wait for someone else to name it.
• Treat structural problems as things to drive, not things to explain. The same observation is a strength when you’re driving it and a liability when you’re naming it as a constraint.
• Build the boss-facing observable-signals scaffolding proactively, before it’s asked for. Make accountability easy upward.
• Have the trust-signal conversation verbally and early, not at performance-review depth.