Decisions that made marketing's numbers smaller and more believable

Retiring lead quotas and letting sales audit the pipeline

In March 2021 Yarrowby measured marketing on marketing-qualified leads. We produced about 14,000 a year, a team of eight sales development representatives called them, and the board saw a marketing-sourced pipeline of A$61 million for the year. Our chief revenue officer told me in my first week that his account executives treated our leads as a chore, and the evidence supported him: about 4 per cent of leads became customers, and acquisition costs were rising faster than new recurring revenue.

The stakes were budget and credibility. Marketing spent A$14 million a year including salaries, and nobody could say with confidence which part of it produced customers who stayed.

I proposed retiring lead quotas. In their place, marketing and sales would agree one definition of a qualified opportunity, which sales would accept or reject, and marketing would be measured on cohort conversion to won revenue, cost per new customer and eighteen-month retention by source. Our head of demand generation disagreed. Her team's bonuses depended on lead volume, and she argued, fairly, that we would lose sight of early demand. Our chief financial officer worried about investors, to whom the company had described its pipeline coverage. I gave up the lead quota, moved the sales development team into sales under the chief revenue officer and accepted that marketing's headline figure would fall.

It fell from A$61 million to A$37 million in the first full quarter. We rebuilt scoring to guide our own programmes rather than trigger calls, asked every new customer how they had first heard of us and published the definition for finance to audit.

Over three years, win rates on qualified opportunities rose from 19 to 27 per cent, cost per new customer fell from A$14,200 to A$10,900 and the months needed to recover acquisition cost fell from 31 to 22. Budget discussions with the board moved from credit to customers.

I would do three things differently. I would run the old lead measure beside the new one for two quarters, so the comparison was clean rather than argued. I would rewrite our investor language with the chief financial officer beforehand rather than leave him to explain the change. And I would find the head of demand generation a role in the new model before announcing it; she left within the year, taking knowledge we needed.

Writing to every customer after a credential-stuffing attack

In May 2023 a payroll officer at an aged-care customer rang our support line: three employees' bank details had changed overnight, and none of them had changed them. Our security team soon established that attackers were trying email addresses and passwords leaked from other websites against our employee app, which then allowed sign-in with a password alone. Over eleven days they had entered 2,140 employee accounts at 58 customers, viewed payslips and changed bank details for 23 people. Nine had a pay redirected before it was stopped.

The technical response belonged to our chief information security officer. What we said, to whom and when was decided between the chief executive, our general counsel and me, and it was contested. The chief executive's first instinct, supported by legal advice about precision, was to write only to the 58 affected customers. I argued that all 7,400 should hear from us within two days, because all their employees used the same app and many would have reused passwords. I also proposed stopping all marketing at once, including an A$2.4 million brand campaign due the following week.

We wrote to every customer within 48 hours, and the board released a statement to the ASX that morning. We repaid the nine employees within two days, notified Australia's privacy regulator and the affected individuals under the Notifiable Data Breaches scheme, and supported our UK customers with their own reports. Multi-factor sign-in became compulsory on every employee account within six weeks, and new bank details were held until an employer approved them. Marketing went silent for eight weeks. The brand campaign was postponed for a year, costing A$610,000 in cancelled bookings and unusable production, and I declined a security-themed campaign that sales wanted three months later.

New-customer pipeline fell 21 per cent for two quarters, support calls tripled during enrolment, and gross revenue retention fell from 95.1 per cent in FY2023 to 92.6 per cent in FY2024. By June 2026 it had recovered to 95.4 per cent.

What I would change begins before the incident. In 2022 I had argued against compulsory multi-factor sign-in because it added friction for shift workers during onboarding. I was wrong, and I told the board so. I would also agree incident letters and decision rights in advance; we wrote ours from scratch at midnight.

Paying accounting partners for customers who stay

By 2022 about 610 accounting practices, bookkeepers and payroll bureaus across our three countries referred employers to Yarrowby. We paid each a one-off fee when a referred customer signed: A$1,500 at home and in New Zealand, £800 in the UK. Partners introduced 22 per cent of our new recurring revenue, and the programme looked healthy until we analysed retention by source. Thirty-eight per cent of partner-referred customers had left within eighteen months, against 14 per cent of those who came to us directly. Many were too small for our product or had been referred by practices that would not help them implement it.

I proposed replacing the fee with a revenue share of 12 per cent of each referred customer's subscription, paid quarterly for three years while the customer remained, and a certified tier for practices that completed our implementation training and supported clients through their first pay runs. Our UK general manager disagreed: the UK business was younger, she argued, its partners had been recruited with upfront fees, and changing them would slow growth just when the region needed it. Our head of partnerships, who had built the original programme, warned that many practices would simply stop referring. Both were right.

We announced the change in August 2022 with four months' notice and honoured the old fee for referrals already in progress. In the first year 190 partners, about 31 per cent, made no further referrals. Partner-referred new customers fell 24 per cent in the second half of FY2023, and the UK business missed its new-revenue plan by A$1.1 million. Over eighteen months we trained 240 practices to the certified standard.

By FY2026, 89 per cent of partner-referred customers were still with us after eighteen months, up from 62 per cent. Partners introduced 34 per cent of new recurring revenue, and certified practices supported 58 per cent of all new implementations. Its annual cost rose, but the cost per customer retained beyond eighteen months fell by about a third.

I would make the change in two steps. Our fifty most productive partners already retained clients well, and they should have helped design the new model rather than learn of it with everyone else. I would also have paid a modest trailing share on earlier referrals, which would have cost little and kept goodwill that took a year to recover.

Repricing payroll without surprising anyone

Until 2025 payroll, rostering, leave and onboarding each carried its own Yarrowby fee per employee per month, with discounts negotiated deal by deal. Customers found their bills hard to predict, and we had 1,900 distinct price combinations across 7,400 customers. Our costs had also risen, particularly for the security and compliance work that followed the 2023 incident. The executive team agreed to move to two per-employee bundles, which would raise the average bill by 8.5 per cent.

The disagreement was about how. Our chief financial officer proposed 60 days' notice, the contractual minimum for most customers, and no cap, so that the full increase reached FY2026 revenue. I proposed 120 days' notice, a cap of 15 per cent on any customer's increase in the first year, a two-year price lock for customers moving to annual billing, and a letter to each customer showing their own current and new monthly bill instead of a link to a pricing page. The chief executive supported my proposal with a condition from the chief financial officer: we would track price-related cancellations monthly and revisit the cap if they ran above our model.

We wrote to customers in February 2025 for a July start. Account managers and partners received each letter a week before the customer did, with the reasoning behind it, and about 400 customers facing uncapped increases above 20 per cent were telephoned first.

The cap and the longer notice cost A$3.6 million of FY2026 revenue against the chief financial officer's plan. About 2,300 customers, 31 per cent of those billed monthly, moved to annual billing to take the price lock, which lowered cash collected in some months while improving retention.

First-year price-related cancellations were 1.4 per cent of recurring revenue, against 3.5 per cent modelled. Net revenue retention rose from 104 per cent in FY2025 to 108 per cent in FY2026, and the number of distinct price combinations fell to fewer than 60.

I would treat our smallest customers differently. Employers with fewer than 60 staff took the largest percentage increases, because the old per-product discounts had favoured them, and they accounted for most of the cancellations. A separate small-employer bundle would have been fairer. I would also brief partners two weeks ahead, not one; several still heard of the change from their own clients first.