Setting the Scene
Every project starts with confidence. At Lumitech — a service IT agency delivering design and software development — the Arq Decisions project began exactly as it should. Arq Decisions — a Hamburg-based decision intelligence platform built to help organizations make faster, more accurate strategic choices — came to Lumitech with a clear scope, a signed-off estimate, and a professional Agile delivery process from day one.
The team ran a thorough discovery phase: stakeholder workshops, client document analysis, and several rounds of requirement refinement. Development was paired with design from day one, so every requirement was backed by prototypes and high-fidelity mockups. Estimates were agreed upon, scope was signed off, and the team moved into active development.
A standard, professional project kickoff. Familiar to anyone in the industry.
Everything Looked Fine — Until It Didn’t
Development followed Agile: sprints, planning sessions, backlog refinement, regular demos. On the surface, the process was running smoothly.
Then the feedback started coming in.
Clients reviewed working features during demos and requested changes
Those changes seemed small — a UI adjustment here, a logic tweak there
But the product was already months into development, with deep interdependencies between modules
Each “minor” fix rippled across multiple areas of the codebase
No single change felt alarming. Cumulatively, they were devastating
All these ‘insignificant’ corrections led the project to a 70% overrun in logged hours. And this is the reality most companies face — because problems only become visible in a report, or worse, in an invoice that now needs explaining.
Chief Product Officer, Tolmete
The Day X Moment
Several months into development, the project manager reviewed actual hours logged against the original estimate.
70% budget overrun confirmed. Not from negligence. Not from poor engineering. From the accumulation of untracked, unscoped change-request-free corrections silently absorbed into sprints.
This is not a Lumitech problem. This is an industry problem:
62% of projects with budget overruns cite scope creep as a primary cause
Average cost overruns from scope creep range from 27% to 45%
(Source: PMI Pulse of the Profession 2023; Productive.io citing IEEE data)
What Happens Next — And Why It’s Almost Always Painful
When a project overruns on a Time & Material contract, there are two possible outcomes:
Client understands the nature of changes
Client understands the nature of changes
They accept the hours
Agency absorbs hours to save the relationship
Margins are protected
Discounted invoices erode profitability
In some cases: end of the partnership entirely
Businesses track costs closely. Not every client can connect “we asked for a small visual update” to “that update required reworking three interconnected modules.”
The gap between technical reality and client perception is where margin goes to die.
Is scope creep already silently draining your project budget?
Tolmete detects unplanned effort, sprint drift, and over-servicing in real time — before the invoice arrives and the conversation gets hard.
How Tolmete Helped After the Fact
When the overrun was confirmed, the Lumitech team used Tolmete’s AI Summary to reconstruct what had happened. Because the team logs time directly in Tolmete, the AI was able to:
Break down hours by feature — identifying exactly where time was spent and which tasks consumed disproportionate effort
Pinpoint the leaks — surfacing which changes arrived without scope approval and when they entered the sprint
Build a clear narrative — giving the PM a factual, structured account of what occurred and why, defensible to the client
Retrospective analysis saved the relationship. But it could not save the margin. This story ends with a recovered client and a hard lesson: the best time to catch scope creep is not after the overrun is confirmed. It’s in the sprint where it starts.
How to Prevent It: Early Warning, Not Post-Mortem
The root cause of the Arq Decisions situation was not the fixes themselves. It was that no one saw them accumulating in real time. Each change request flew under the radar because individually, none crossed a threshold that triggered a conversation.
Tolmete connects to your existing tools — Jira, GitHub, Slack — and continuously analyzes team activity against project scope. Instead of discovering problems at milestone reviews or invoice time, you receive signals while there is still room to act.
Signals Tolmete surfaces early include:
Scope Creep — unplanned work entering sprints without change requests, flagged as it forms
High Churn Rate — code being repeatedly rewritten, signaling unclear requirements before they consume the budget
Unverified Activity — work happening outside tracked tasks and tickets, creating invoice risk before the client sees a number
Bus Factor — critical knowledge concentrated in one team member, a delivery risk hiding in plain sight
Task Stagnation — tickets sitting in “In Progress” far longer than comparable work, silently draining capacity
A real Tolmete signal on a live sprint:
SIGNAL: Unverified Activity
The unplanned work likely originated from direct client communication. Marcus's timelog references a client call, and the webhook logging PR was motivated by a client-reported issue. This work was absorbed into the sprint without adjusting the commitment or filing change requests. On a sprint already tracking at 53% completion on day 7, 14 additional hours of unplanned work puts the sprint at significant risk of failure.
Notice what this is not: a generic alert. It identifies who, why, when, and what the consequence will be — while there is still time to change the outcome.
From Signal to Action: The AI Mitigation Plan
Tolmete does not stop at identifying the problem. Every signal is accompanied by an AI Mitigation Plan — specific, directed, actionable steps assigned to the right people:
Immediate Actions
PM → Review the 3 untracked PRs with Marcus and Alex today. If client-driven, file formal change requests retroactively to protect project scope.
Marcus Reed + Alex Kim → Create Jira tickets for all 3 PRs retroactively and link them properly. The project record must reflect what actually happened — this protects billing accuracy and preserves sprint history.
This Sprint
PM → Reassess Sprint 22 delivery forecast. With 14 hours of unplanned work absorbed and only 53% completion on day 7, decide now whether to drop lower-priority tickets or communicate a revised delivery date.
Next Sprint
PM + Tech Lead → Establish a team agreement: any out-of-sprint client request requires a Jira ticket before work begins, even for urgent fixes. This doesn’t slow down urgent fixes — it makes them visible.
The Arq Decisions story is not about a team making mistakes. It is about a fundamentally reactive management model — where problems are only visible once they have already compounded.
The difference between a disputed invoice and a healthy client relationship often comes down to one thing: how early you knew.
With Tolmete:
Your project is under continuous monitoring across all connected tools — not just at end-of-sprint review
Problems surface during the sprint, not after the invoice is sent and the damage is done
Your team gets directed, actionable plans — not generic recommendations, but named owners and specific next steps
You protect margins proactively — rather than defending them retrospectively in front of a skeptical client
Tolmete is an AI delivery intelligence system for service companies.
Connect your tools, monitor your projects, and give your team the time to act before small issues become expensive problems.
FAQ
Frequently Asked Questions
Most budget overruns aren't caused by poor performance — they're caused by invisible scope growth. Client feedback, mid-sprint "quick fixes," and demo revisions rarely trigger formal change requests. Each one seems too small to escalate. But across a multi-month project, these micro-additions compound into hours that were never scoped, never priced, and never approved. By the time the overrun appears in a report, the budget is already gone.
Traditional reports describe what already happened. Scope creep detection warns about what is forming right now. A Jira export or time tracking report shows logged hours after a sprint closes — it won't tell you that 14 unplanned hours entered on day 3, or that a "minor fix" triggered rework across three modules. The practical difference: one tool produces a post-mortem, the other prevents it.
Margin protection and micromanagement are not the same thing. Micromanagement means asking people what they're doing. Margin protection means understanding whether delivery is drifting from what was sold. A system like Tolmete works on metadata only — commit frequencies, ticket transitions, sprint patterns — never private messages or individual activity. The result: PMs spend less time chasing status updates and more time acting on early signals, which is exactly what high-performing teams respond well to.
