Case study: a 12-person SaaS startup moves from Trello to Microsoft 365

Names and identifying details have been removed at the client’s request. The pattern below is one we see often enough that the lessons travel even when the specifics don’t.

The client

The client is a North American SaaS startup with roughly 12 people on the team — engineering, design, sales, and operations — building project management software for a mid-market customer base. Post-seed, growing, the kind of company where the product is good, the market is real, and the operational tooling has fallen quietly behind. They came to us at the point where “we’ll figure it out as we go” had stopped working: the team was effective but the seams were starting to show, and an upcoming Series A diligence pass was about to make those seams everyone’s problem.

The problem

The pain points were familiar and connected:

• A free Trello board had been the system of record since founding. It hit the free tier’s limits about a quarter before our first call, and the paid tier didn’t actually solve their workflow — most of the friction was in how the boards had grown, not in licence limits.
• Email lived in personal Gmail accounts mixed with founder-name domains. There was no centralised inbox, no tenant, no MFA enforcement, and nothing that would survive a founder leaving.
• Documents were strewn across Dropbox, Google Drive, and local folders. The same contract sometimes existed in three places with three slightly different versions.
• Technical specs, customer onboarding plans, and signed contracts had no canonical home. Finding a year-old onboarding deck meant asking three different people on Slack and hoping one of them remembered who’d written it.
• The lead investor’s diligence team flagged the document trail as something to clean up before the round closed.
• They needed to scale from 12 to 25-plus headcount over the next 18 months without any of these problems compounding.

Nothing on this list is unusual for a 12-person startup. It just doesn’t survive contact with growth or a serious investor review.

What we recommended

We started where we always start — a free assessment plus a short paid discovery to scope the actual work. The deliverable was a written report and a fixed-fee proposal for the migration phase, with the optional next step of an ongoing pay-as-you-go support arrangement.

The recommended target stack:

• Microsoft 365 with mixed licensing. Business Premium for the two founders (managed laptops, sensitive financial and HR data, MFA enforced through Conditional Access). Business Standard for the rest of the team. Premium for everyone would have meant a meaningful additional spend [verify ~$X CAD per year for this scenario] with no functional gain for the engineering and sales staff who didn’t need Intune or Defender for Business.
• SharePoint as the document hub, with information architecture designed before any sites got created. Site collections by function — Engineering, Sales, Operations, Legal — each with permission boundaries that matched how the team actually worked, not how the org chart looked.
• Teams for chat, meetings, and channel-based work. The existing fragmentation across personal accounts, group DMs, and external invites went away as part of cutover.
• Project Online for portfolio and resource tracking. Slightly heavyweight for the current size, but specifically chosen because the client wanted to dogfood Microsoft’s PM tooling alongside their own product.
• Conditional Access policies as a security baseline — MFA, geographic restrictions, and device compliance rules where Premium licences applied.
• A staged migration over three consecutive weekends so the workweek was never disrupted: email cutover first, document migration second, Teams and Project Online go-live third.

What we did

The actual delivery looked like this:

Tenant setup with security baselines. New Microsoft 365 tenant on the client’s primary domain, MFA enforced from day one, six Conditional Access policies covering the obvious cases (block legacy auth, require MFA, restrict admin roles to known geographies, require compliant device for Premium-licensed users), and break-glass admin accounts documented and stored in the founders’ password vault.

SharePoint architecture. Four site collections — Engineering, Sales, Operations, Legal — designed around real permission boundaries rather than reporting lines. Engineering had open membership across the team for shared technical material plus a private subspace for sensitive artefacts; Sales had per-deal channels; Operations held HR, finance, and vendor docs; Legal stayed restricted to the founders.

Document migration. Roughly 8,000 files moved from Dropbox and Google Drive into the new SharePoint structure, with metadata preserved where it existed and applied retroactively where it didn’t. A specialised migration tool handled the bulk move; we did the manual reorganisation that no tool can automate.

Email cutover. Personal Gmail mailboxes forwarded into the new Exchange Online tenant for two weeks while users got comfortable, then a hard cutover at the end of the second weekend. Existing-thread continuity was preserved by importing the prior 24 months from each mailbox.

Teams rollout. Channel structure mirrored the SharePoint sites so links and files lived in one place rather than two. The previous group DMs and external invite history were intentionally not migrated — that conversation history was worth letting fade.

Project Online configuration. A single project portfolio with custom fields for the client’s two-track use case: internal project tracking on one side, structured product feedback from their own customer base on the other.

Training and handoff. Two hours of role-based training per person, recorded so new hires could watch later. Written documentation of every config decision, an admin runbook, and a half-day handoff session with the operations lead so they could maintain it themselves.

Outcome

Six weeks from kickoff to go-live. No downtime during cutover. Concrete numbers reported back at the 90-day check-in:

• Document search time dropped from “days” — the client’s own word, when finding a year-old onboarding doc meant asking three people on Slack — to seconds against indexed SharePoint search.
• Investor due diligence completed in roughly two weeks versus the six the lead investor’s team had originally estimated. The Engineering and Legal site structures held up under direct review without anyone scrambling to assemble files.
• Headcount grew from 12 to 18 over the following year. Onboarding a new hire — accounts, licences, training, full SharePoint and Teams access — became a 30-minute administrative task. Previously it had been a two-day scramble nobody owned.
• Total Microsoft 365 spend averaged roughly [verify ~$X CAD per user per month] across the mixed-licence stack, materially below the all-Premium price tag the founders had originally assumed.
• Ongoing support continued through our pay-as-you-go arrangement: a small monthly minimum keeps Embrollar on the roster, billable hours when the client actually needs us. After the first quarter, monthly support hours have stayed low and predictable.

What we’d do differently

Two things, in honest reflection:

Training time was underestimated. We scoped two hours per person; SharePoint navigation alone needed three. Most of the questions in the first month were “I can’t find X” — solvable with better up-front orientation rather than reactive support tickets.

Project Online was the right tool eventually, the wrong tool first. The learning curve was steeper than the team had bandwidth for in the first six months. Microsoft Planner would have done the job during the early phase; next time we’d recommend Planner first and introduce Project Online once the portfolio actually warranted the heavier tool.

Hitting the wall on light tools?

If your business is at the point where free or light tools have stopped working — too many docs in too many places, no central tenant, an investor or auditor about to ask hard questions — that’s exactly the situation our free assessment is designed for. We will tell you straight whether you need an engagement like the one above, a smaller fix, or nothing at all yet. Same pay-as-you-go model afterwards: a small monthly minimum, billable hours only when you need us. We don’t put you on a retainer that doesn’t fit your stage.