Onboarding That Actually Works: How I Took a 90-Day Client Onboarding Disaster Down to Under 30
Twelve clients stuck in an onboarding backlog. Three already gone, having terminated their agreements before the process even finished. And a leadership team asking me if I could step in and fix it.
This is the story of how a broken managed IT services onboarding went from 90+ days of chaos to under 30 days of structure. The fastest one we completed: just over two weeks.
Situation
A mid-market managed IT services firm supporting roughly 75 clients and 3,000 end users. Every time a new client signed an agreement, the service transition team had to deploy the full tech stack (endpoint protection, monitoring tools, backup systems, security awareness platforms) while simultaneously building out all the documentation, runbooks, and resources the support desk and account managers would use to deliver ongoing services.
Under normal conditions, the team onboarded one to two new clients per month. The work was complex, cross-functional, and detail-heavy. Done well, it set the foundation for every interaction the client would have for years. Done poorly, it set the tone for a relationship that started underwater.
It was being done poorly.
Challenge
When I stepped in, the onboarding process had ballooned past 90 days. In some cases, well beyond that. Twelve clients were sitting in an active backlog, in various states of progress, with no clear picture of what was done and what was still outstanding.
The problems ran deep.
No structure. There was no centralized tracker, no task management system, no way for anyone, including the team doing the work, to see what had been completed versus what was still in flight. Work was tracked in people's heads and scattered across emails.
Communication breakdown. Internal teams were asking the same questions through direct messages, getting different answers depending on who responded. Clients were left in the dark. Account managers didn't get involved until the tail end and had no context on what happened during the transition. Sales handed off the deal and disappeared.
Resource bottleneck. The service transition team was three people. When workload spiked or someone was out, everything stalled. The backlog grew. Tasks got missed. The same work got done twice. Or not at all.
Clients were walking away. This was the part that got leadership's attention. Three clients terminated their agreements before onboarding was even completed. Others said it directly: if things don't turn around, we're canceling.
That's when the Senior Leadership team asked me to step in.
Action
The problem was obvious on the surface: everything was taking too long. But you can't fix "too slow" without understanding why it's slow. I needed to see the process from the inside before I could rebuild it.
Week 1: Triage
The first thing I did was set up three 15-minute meetings a day with the service transition team. Morning, midday, and late afternoon.
The purpose wasn't status updates. It was prioritization. At each meeting, the team had to tell me what they were working on and why. I'd listen, ask questions, and then tell them the one thing they should be focused on until the next meeting. The goal was to stop the team from spreading across ten tasks at once and instead finish one thing at a time.
During those meetings, I started mapping the process. Every task, every dependency, every handoff point. By the end of week one, I had a working picture of what needed to happen in an onboarding, and where it was breaking down.
Weeks 2–8: Structure
With the process mapped, I started building the system.
Tracking. I started with a simple Excel spreadsheet so everyone (not just the onboarding team, but anyone in the company) could see where each client stood: tasks completed, in progress, and not yet started. That spreadsheet eventually moved to a Kanban board in Microsoft Planner.
Prioritization. I ran every task through the Eisenhower Matrix. Urgent items got sequenced by priority and deadline. Important-but-not-urgent items went to the backlog. Everything else got cut or pushed to post-onboarding. This killed the "do everything at once" pattern that was stalling the team.
Communication. I announced that all onboarding-related communication was moving to a dedicated Microsoft Teams channel. No more DMs to me or anyone on the team. If someone had a question, it went to the channel. Because if one person's asking, others need the answer too. It also meant nobody was answering the same question multiple times.
Handoff gaps. This was the hardest piece. Using journey mapping, I traced the full client lifecycle from signed agreement to go-live and identified two major gaps: the handoff from sales into transition, and the handoff from transition into account management.
I brought both teams together, showed them the gaps, and we built the fix collaboratively. The biggest change: I added mandatory meetings at the start and end of the process. The kick-off call required both the salesperson and the assigned account manager. If either one couldn't attend, the meeting didn't happen. It got rescheduled. No exceptions. This created a true relationship hand-off: the client met their AM on day one, the AM had full context, and the salesperson could step back knowing the relationship was covered.
The reframe. The single most impactful change was philosophical. The old process aimed for perfection: every documentation page complete, every system fully configured, every last detail buttoned up before the client was considered "live." It was thorough on paper. In practice, it took forever.
I reframed the entire process around speed to value. I asked one question: what are the two things a new client actually cares about? The answer was simple. They want help from the support desk, and they want to start working with their account manager as soon as possible.
Everything that didn't directly enable those two things got stripped out of onboarding and moved to ongoing service delivery.
The new process had two phases:
- Phase 1 (~2 weeks): Speed to support. Deploy the minimum tech stack and documentation needed for the support desk to take calls and resolve tickets for the client.
- Phase 2 (~2 more weeks): Full go-live. Complete the remaining configurations, finalize account management resources, and officially hand off to steady-state service delivery.
That was it. No ten-phase project plan. No perfection. Just the two things the client was paying for, delivered fast.
Weeks 8–12+: Scale
With the process validated, I made the structural change that would make it sustainable. Instead of a dedicated three-person team doing all the onboarding work, I shifted the model to a single process manager supported by the full service delivery team.
All tasks were documented so any technician, from tier one through tier three, could pick them up based on skill level. Tickets were created, queued, and assigned with due dates. The process manager's only job was to keep the system moving: make sure tickets were completed on time, escalate blockers, and continuously improve the process.
It was identical to how the support desk already worked. The only difference was that these were onboarding tickets instead of support tickets. The team already had the muscle memory. They just needed the structure.
Result
Within 12 weeks, the new process was validated. The first clients taken through it completed onboarding in under 30 days. The fastest: just over two weeks.
Here are the results that mattered:
- 90+ days → under 30 — with a fastest completion just over 2 weeks
- Zero client churn during onboarding — after the fix, not one client terminated during transition
- 3 at-risk clients saved — all three later confirmed directly that they would have canceled without the turnaround
- 18 onboardings completed — the original backlog of 12, plus 6 new clients added during the fix period
- 3-person team → 1-person process role — shifted from a dedicated team to a single process manager leveraging the existing service delivery team
- Sales and AM engagement fixed — both teams reported significant improvement in their experience with the onboarding process
The director of managed services credited the work with preventing further client losses. Three clients specifically said, unprompted, that they would have terminated their agreements without the changes.
Takeaway
The instinct when something is broken and slow is to add more. More people, more steps, more checkpoints. The fix here was the opposite. It was subtraction.
The old process failed because it tried to do everything during onboarding. Every document. Every configuration. Every detail. The problem wasn't that people weren't working hard enough. It was that the process itself was designed around internal completeness instead of client value.
Speed to value over perfection. That was the entire shift. Strip the process to the two things the client actually cares about, deliver those fast, and move everything else downstream.
One honest reflection: I should have taken one more week to assess before moving at full speed. The urgency of the backlog pushed me to act fast, and there were a couple of early client experience hiccups that an extra week of diagnosis probably would have caught. In the end, they were minor. But "move fast" has a cost, and being honest about that cost is part of what makes an operator credible.
If your onboarding process takes too long and your clients feel it, the answer probably isn't a better project plan. It's a harder question: what can you stop doing?