Chief Human Resources Officer / Canada
Hard decisions, with the arithmetic shown
- Colleagues in four offices and three countries
- 1,310
- Kestrelford Software Inc., June 2026; peak of 1,520 in 2022
- Annual compensation and benefits expense
- C$238m
- Company-wide, 2025; not the HR operating budget
- Eligible colleagues who accepted the 2024 option repricing
- 93%
- 571 of 612 underwater option holders
Decisions & consequences
01Cutting 214 roles by stopping three pieces of work
By December 2022, after eighteen months of hiring from 1,140 to 1,520 people, Kestrelford's growth had slowed to a forecast 22%, and the board wanted positive free cash flow by early 2025. Our chief financial officer, Marisol Fennick, proposed reducing every department by 22%, about 330 roles, in one step. Ottilie Brennagh, who chairs our people and compensation committee, favoured a single deep cut so that nobody would live through a second.
I argued for a different starting point, having carried out two even cuts at Tamberlane Systems and watched the second follow from the first. I asked the executive team to name the work we would stop. We ended a rewrite of our field inspection app, a direct-sales expansion in Australia and an internal analytics platform, and froze hiring for six months. That removed 214 roles, 14% of the company, and pushed positive free cash flow about two quarters beyond the finance plan. Our chief executive wanted the announcement to promise no further round. I would not write that, and we stated instead what would trigger another review.
Each affected person heard from their own manager in a scheduled call within one four-hour window, then from HR, and kept system access until evening and their laptop for good. Severance was twelve weeks plus three per year of service, with six months of benefits, and vested options could be exercised for three years instead of ninety days. With more than fifty Waterloo colleagues affected, we gave Ontario its required mass-termination notice. The 23 colleagues on employer-specific work permits could choose working notice to stay in status longer.
One-time costs were C$27.8 million and the run-rate saving C$38 million. Free cash flow turned positive in the third quarter of 2025, and there has been no further reduction. Of the 214, 71% told us they had new roles within four months. Agreement that leadership could be trusted fell from 74% to 49% and recovered to 68% by September 2025.
What would I do differently? Counsel found nine people on parental or medical leave among 58 selections within continuing teams, and five of those rested on recent visible output. We reversed them, but I should have required written role criteria before managers chose. I would also fix the announcement date sooner: nineteen days of rumour cost more trust than the decision.
Repricing underwater options after the April 2024 financing
Kestrelford's Series F in April 2024 raised C$150 million at a post-money valuation of C$1.9 billion, against C$3.1 billion in 2021. The third-party valuation of our common shares fell from C$13.40 to C$5.90. Of the 1,260 people holding options, 612 held grants made between 2020 and 2023 with exercise prices above the new value, many of them engineers for whom equity was a fifth or more of annual pay. Competitors from the United States were calling them weekly.
There was a second problem. Most employees had read each valuation as the worth of their options, unaware that C$720 million of investor preferences ranked ahead of common shares. Repricing without explaining that would repeat the mistake.
I proposed resetting every underwater option to C$5.90 in exchange for adding twelve months to its remaining vesting, excluding the fourteen members of the executive team for a year, and sending every option holder a statement showing their holding's value after preferences at four exit values, one of them below the preference total. Callum Thistlewood, our chief executive, argued that executives should be included because they were the people competitors most wanted. Ottilie Brennagh questioned whether resetting prices rewarded employees for a lower valuation, and asked for the incremental accounting cost. I told the committee that including executives would make the programme look like a management benefit, and that the statement mattered more than the price.
The committee approved the design. We held forty-one sessions in English and French, paid for individual tax advice in three countries and gave people six weeks to decide. 571 of the 612 eligible colleagues accepted. Regretted attrition among engineers fell from 10.9% in the twelve months before to 6.2% in the twelve months after. The incremental share-based compensation expense is C$7.4 million over three years. Two of the fourteen excluded executives left within the year; both cited the exclusion. Executives became eligible in 2025 under a smaller reset.
The annual statement is now part of every option holder's year. What I would do differently is the vesting extension. A flat twelve months penalised people close to full vesting, and most of the 41 who declined were in that group. I would scale the extension to the vesting remaining.
Publishing every pay band to every employee
In early 2025 we began preparing for Ontario's requirement, from January 2026, to include expected pay or a range in publicly advertised postings. We already posted ranges in British Columbia and in the US states that required them, so the external change was modest. The internal one was not. When we compared every employee's pay with the band structure we intended to publish, 187 people were below their band minimum and 52 were above the maximum. Colleagues hired in 2021 and 2022 earned on average 9% more than earlier hires at the same level. Once ranges were public, anyone could compare their pay with a stranger's advertisement.
Marisol Fennick, our chief financial officer, recommended posting ranges only where the law required and correcting below-minimum pay over two years. Our head of sales objected to publishing ranges for quota-carrying roles at all. I recommended publishing the full band structure to every employee in September 2025, three months before the Ontario rules applied, and moving everyone below minimum before publication. That cost C$3.9 million at once. To fund part of it, the 2026 merit budget fell from 3.5% to 3.0% for everyone, which I told employees directly. People above the maximum kept their pay, but their increases were limited to lump sums until the band caught up.
We gave up managers' freedom to set pay inside a wide undisclosed range, half a point of merit increase for the whole company and the goodwill of some of the 52. Seven of them left within the year. Sales ranges were published with the split between base pay and planned commission at full quota shown separately.
By June 2026, offer acceptance had risen from 76% to 85%, and agreement with the survey statement that people understood how their pay was set rose from 38% to 71%. The unexplained gap between women's and men's pay within the same level and job family fell from 3.8% to 1.1%. The band structure now also anchors our Québec pay equity audit.
What would I do differently? I trained managers in the week of publication, and in the first fortnight HR received more than 600 individual questions, many of which a prepared manager could have answered. I would train managers six weeks ahead and give each one their team's position against the bands before anyone else saw it.
Keeping co-op and graduate hiring when AI tools arrived
In November 2024, during planning for 2025, our engineering leadership proposed reducing co-op work terms from 96 a year to 30 and graduate hires from 34 to 8. Kasimir Lisowski, our chief technology officer, cited a pilot in which experienced engineers using AI coding assistants finished routine changes about a quarter faster. He proposed moving the C$4.6 million saved into fourteen senior hires. The argument was serious: juniors had spent much of their first year on exactly the routine changes the tools now accelerated.
My concern was the senior bench of 2030. Almost half our current senior engineers had joined us as co-op students or graduates, and the Waterloo co-op system works on relationships that take years to rebuild once a company withdraws. I also doubted that the pilot measured the whole job. It counted routine changes completed, not how many of them a customer later reported as defects.
I proposed keeping 72 co-op terms and 22 graduate hires and redesigning what juniors do. In the new model, first-year engineers write the tests and specifications that AI-assisted changes are checked against, review generated code alongside a senior and own small features from specification to release. Each spends two weeks with our support team in Halifax, answering customer tickets. Mentors have 10% of their time recorded in their objectives. Kasimir disagreed in writing and his memo went to Callum Thistlewood with mine. Callum accepted my proposal for 2025, with a review after twelve months, and we deferred six of the fourteen senior hires, about C$2.9 million of planned cost.
After eighteen months, time from a graduate's start date to their first independent production change had fallen from eleven weeks to six. Co-op students accepting our return offers rose from 58% to 74%, and first-year retention of graduates was 95%. Production incidents traced to changes by first-year engineers were no more frequent than the company average. Kasimir now presents the junior programme to the committee himself.
What would I do differently? I scheduled the first support rotations in the weeks before a major release, and several juniors read the timing as a sign the rotation mattered less than delivery. I would set rotations in quieter periods and pay travel to Halifax from the start, and include Montréal in the first year rather than the second.