Summary
Our core market, preschools in Singapore, had become saturated, and I was responsible for growing revenue. I led the strategy and execution of the company's expansion into a new vertical market, children and youth social services. I shaped the product with 6 organisations as design partners, launched with 2 anchor organisations as reference customers, and brought in $1M in new revenue.
As COO, I led a 30+ person cross-functional team and owned the market analysis, discovery, PRD, tender submissions, security and data protection reviews, sales, deployment and change management.
Why did we need a new market?
Cyberland is a Singapore software company whose core market is preschools and kindergartens. Centres use our platform to manage enrolments, track child development and talk to parents.
Several factors showed this market had stopped growing, so we had to pivot:
- Singapore's resident total fertility rate (the average number of children per woman) fell from 1.10 in 2020 to 0.97 in 2023, well below the replacement level of 2.1 and among the lowest in the world
- Resident births fell 17.2%, from 36,875 in 2016 to 30,518 in 2023
- Government-supported preschools, which charge lower fees and run their own internal software, were expanding and shrinking the market we could sell into.
Sources: Singapore Department of Statistics; ICA Report on Registration of Births and Deaths.
Where should we go next?
The question was: which new market is worth winning, and can we enter it with what we already have? With existing clients to serve and limited resources, I compared 5 candidate verticals.
I rated each vertical High, Medium or Low on 5 criteria, then turned the ratings into 2 scores:
| Score | Built from | How it's calculated |
|---|---|---|
| Worth winning | Revenue potential, competition | Average of the 2 ratings. This mattered most. |
| Easy to enter | Domain knowledge transfer, platform reusability, brand transferability | Average of the 3 ratings |
Ratings convert to points: High = 3, Medium = 2, Low = 1. Competition is reversed, because less competition is better: Low = 3, High = 1. For example, Aged Care scores 3.0 on worth winning, (3 + 3) ÷ 2, but only 1.7 on easy to enter, (2 + 2 + 1) ÷ 3.
| Vertical | Domain Knowledge Transfer | Platform Reusability | Brand Transferability | Revenue Potential | Competition |
|---|---|---|---|---|---|
| Aged Care | Medium | Medium | Low | High | Low |
| International Schools | Low | Low | Medium | High | High |
| Children & Youth Services | High | High | High | High | Low |
| Student Care | High | High | High | Medium | Medium |
| Tuition Centres | Medium | Medium | Low | Medium | High |
Children and Youth Services
Worth 3.0 · Ease 3.0
Aged Care
Worth 3.0 · Ease 1.7
Student Care
Worth 2.0 · Ease 3.0
International Schools
Worth 2.0 · Ease 1.3
Tuition Centres
Worth 1.5 · Ease 1.7
Why Children and Youth Services was the clear winner:
- Digitisation mandate. Government funders were actively pushing Children & Youth Services organisations away from paper-based processes. Compliance reporting requirements were growing faster than spreadsheets could absorb.
- Active procurement budgets. Government-affiliated bodies had digitisation budgets and were actively seeking vendors. The willingness to pay existed. The right product just hadn't shown up yet.
- Greenfield market. The CRM tools these organisations were using were free: nonprofit arms of larger vendors, provided as goodwill with no meaningful support or roadmap. Any changes required budget that either wasn't approved or didn't exist. The result was a sector stuck on tools that nobody was improving, with organisations that had no path to upgrade on their own.
With the vertical confirmed, I moved into discovery to pressure-test the assumptions before writing any of the PRD.
What was really broken for these organisations?
I ran discovery interviews and design thinking workshops with 6 organisations across 3 care categories: children's homes, special needs schools, and early intervention centres. We reached them through referrals from existing clients. In each session, operational staff and directors mapped how work moved through their organisation and where it broke down.
The same 3 problems surfaced in every session, without prompting.
- Generic CRMs couldn't fit their needs. Each organisation runs casework its own way, and no off-the-shelf tool handled that variety.
- No shared record of each client. Client information sat in personal Excel sheets and Word documents, so teams working with the same client couldn't see the full picture or plan the next step together.
- Audit season was dreaded. Audits by Singapore's Ministry of Social and Family Development (MSF), the National Council of Social Service (NCSS) and funders meant rebuilding records from scratch, often over public holidays.
Root cause
Scattered records
Client files spread across personal Excel sheets and Word documents
Which caused
Unreliable case management
No team could see a client's full history or coordinate the next step
Which caused
Painful audits
Every audit meant rebuilding records from scratch, often over public holidays
What did we build first?
What made v1
Every scoping decision came down to 2 criteria:
- Is it part of the critical workflow?
- How much impact does it have, in time saved and the number of people affected?
Features on the critical workflow with the highest impact went into v1. Everything else was documented and deferred.
| Feature | Priority | Rationale |
|---|---|---|
| Enrolment | v1 | Entry point for every client record. Nothing else works without it. |
| Case Management Module | v1 | Centralises records across departments. Solves the core visibility problem. |
| Audit Trail | v1 | Timestamps every action. Directly addresses accountability and compliance gaps. |
| Report Generator | v1 | Built on clean centralised data. Generates audit reports from records staff already keep, instead of rebuilding them from scattered files. |
| Transport Management | Deferred | Operational need but not the reason organisations were switching platforms. |
| Billing | Deferred | Real requirement, low discovery priority. Queued for next phase. |
| Scheduling | Deferred | Coordination need identified in discovery but not universally critical across all organisation types. |
| Non-Essential Case Reports | Deferred | Edge-case compliance documents. Not universal across organisations. |
Configure, don't rebuild
Engineering estimated that 70% of what these organisations needed already existed in our platform. Because it transferred so well, we configured it for the new vertical instead of rebuilding, and launched v1 within 6 months. I wrote the PRD and took it to our engineering leads and CEO for resourcing sign-off.
How did we win the first customers?
Starting with early intervention
We started with early intervention centres. They serve children aged 0–6, the same age group our preschool platform already served, so our referral network carried over from day 1.
Anchor customers, chosen on purpose
- Why anchors: we targeted 2 large, well-known organisations, because the first to sign would be the one every other organisation looked to. As a new name in the sector, we needed that credibility more than a quick first deal.
- The trade-off: a longer enterprise sales cycle. I ran the tender submissions, security architecture reviews and data protection checks under Singapore's Personal Data Protection Act (PDPA).
- Keeping revenue moving: we closed smaller deals in parallel, so revenue kept coming in.
- The payoff: once both were live, they became the reference case for the sector, and selling to other organisations got much easier.
Caseworkers first, then leadership
- Operations first. Every engagement started with caseworkers, who spent 20–30% of their time on manual compliance reporting. Every demo led with a before-and-after of their working week.
- IT and leadership second. Once caseworkers were convinced, we brought in IT and leadership for security, PDPA and data governance review.
Caseworker buy-in gave leadership a demand signal to respond to, not a cold pitch to evaluate.
How did we get caseworkers to use it?
Signing the contract was only the start. We rolled out in 5 steps:
Step 1
UAT with key stakeholders
Gaps fixed before any end user touched it
Step 2
Training by department
Each team on its own modules and workflows
Step 3
Champions in every team
Trained to support their colleagues
Step 4
Phased go-live
Core operations teams first, other modules once stable
Step 5
3 months of intensive support
Issues caught before they became habits
What counted as done
We tracked renewal rate, satisfaction and time saved on compliance reporting. All of them trace back to whether caseworkers used the platform in their daily work. We set the bar before launch: the rollout wasn't done until at least 80% of caseworkers rated the system 4 stars or above.
Did it work?
Commercial
- Generated $1M in total new revenue from the new vertical
- Both clients renewed and expanded into adjacent services including special needs schools and other social service units
Credibility and Market Position
- Both anchor organisations were recognised names in Singapore's youth and social service sector
- Their early adoption established Cyberland as a credible vendor in a referral-driven market, directly unlocking the pipeline that made the $1M target achievable
Product and Operational
- A post-rollout survey of 120 caseworkers showed they saved an average of 1 hour per day previously lost to manual compliance reporting, and the rollout cleared our bar of at least 80% rating it 4 stars or above
Strategic
- 70% platform reuse meant Cyberland entered a new vertical without a ground-up rebuild, preserving margin while proving the core CRM architecture was transferable across adjacent sectors
What would I do differently?
Change management takes longer than the product does
- Mistake: Underestimated the inertia of non-tech-savvy users shifting to a new system. Training users only at go-live wasn't enough to drive independent adoption.
- Lesson: Build sustained, repeated training into the post-launch phase from day 1: not just availability, but structured sessions until users operate without supervision.
Stickiness needs more than 1 champion
- Mistake: Each organisation started with 1 main champion, and we relied on that person to drive adoption.
- Lesson: Start with a main champion, but deliberately grow more champions across departments over time, so adoption doesn't depend on any single person and the product stays sticky.
Enterprise deals take longer than the build
- Mistake: Underestimated how much time tenders, security reviews and data protection checks add to enterprise deals in government-funded sectors.
- Lesson: Plan team capacity and timelines around the full sales cycle, not the product timeline. In regulated sectors, assume each deal takes 2x as long as expected.
UAT doesn't replace real-world validation
- Mistake: UAT and sign-off caught what users knew to test for, but edge cases only surfaced once the platform was live at scale across 50–300 caseworkers.
- Lesson: Run a structured 30-day parallel trial alongside the existing system before full go-live. Real daily use surfaces what controlled testing never will, and makes the eventual launch far cleaner.