Most software ROI calculators pick the most generous assumption available, multiply by your team size, and produce a number that says you'll save $4M. We've gone the other way.
This post walks through the ROI model we build for every prospect, what's in it, what the anchors are, and which inputs we've deliberately kept conservative so the result is something a CFO can defend rather than a number you have to apologise for.
The anchor: 14 hours per week
There's one published number we trust as the foundation. The 2018 Construction Disconnected report - a survey of 600+ construction leaders worldwide, run by PlanGrid (now Autodesk) and FMI - found that engineering and construction professionals spend roughly 14 hours per week, about 35% of their time, on non-productive work: looking for project information, resolving conflicts, dealing with mistakes and rework.
That's the only fixed anchor in our model. Everything else is conservatively layered on top.
Value driver 1: Time savings
The simplest driver. CalcTree gives back a portion of those 14 hours through single source of truth, version control, AI-assisted review, and reusable templates that replace ad-hoc Excel files.
We don't claim to capture all of it. The default capture rate is 15%, meaning we model CalcTree as reclaiming about 2 hours per user per week of the 14hrs of waste. Most teams will do better than that. We use 15% based on initial pilots with user groups and 4 years of research to design CalcTree.
The formula:
hours saved per user per month = 14 hrs × capture_rate × 4.33 weeks
$ value per user per month = hours saved × hourly rate
For a typical engineering team at a fully-loaded hourly rate of around $150, that's roughly $1,400 per user per month in time value.
Value driver 2: Risk reduction
The harder thing to quantify, but real. CalcTree's reviews, version control, and single source of truth catch errors that otherwise become rework, change orders, or schedule slips. The same FMI/Autodesk research puts US construction rework at $177B annually, 70% traceable to design errors.
To turn that industry-level figure into a per-user-per-month number without overclaiming, we go bottom-up:
- Cost of one prevented event. A meaningful calc or spec error caught before it propagates downstream. Conservative average: $10,000 (rework hours and minor schedule impact). Range: $5K for a small fix, $100K+ for a major one.
- Frequency. For a 3-person team, conservatively one prevented event per year, team-wide. Reasonable case: one per quarter.
- Spread across team and time: $10,000 ÷ 3 users ÷ 12 months ≈ $278/user/month, rounded to $300.
There's no public benchmark for "per-engineer-month risk reduction." The right thing to do is to ground it in your historical incident rate during the pilot. Get in touch to get a copy of our interactive calcultor to plugin your own numbers here.
Value driver 3: Compounding upside
CalcTree gets more valuable per user as more colleagues join; shared templates, fewer "which version is right" disputes, easier collaboration on calcs. We model this with a capped multiplier so the projection doesn't run away.
The bonus grows slowly and caps at around 5% in our conservative default. Even with hundreds of users, it only adds a few percent per person. this is deliberately modest.
Two scenarios, side-by-side
Every model we share shows two cases:
- Growth scenario — users expand over 36 months (e.g. 3 → 35)
- Floor scenario — users stay at the pilot size forever
The floor matters because it answers the obvious sceptic's question: what if rollout stalls? If a model only works under aggressive growth, it isn't defensible. Ours holds even at the floor, typically still a positive return inside the pilot window.
What a typical model looks like
For a 3-seat pilot expanding to ~35 users over three years, with conservative settings:
- 3-year value: $1.0M–$1.3M
- 3-year cost: $30K–$40K (CalcTree subscription)
- Payback: typically within the pilot window
- 3-year ROI: 25–35×
The range reflects what you'd plug in for hourly rate, growth path, and risk frequency. Your champion can flex any of these in the live model.
Three inputs to customise for your team
Three of the defaults are worth replacing with your own numbers:
- Risk reduction value. Our $300/user/month is a starting point. Replace with a figure grounded in your incident history — cost of a typical prevented event × how often you'd realistically prevent one. Your data beats our default every time.
- Capture rate. We default to 15%, deliberately conservative. If your team can name specific time-sinks CalcTree directly addresses — chasing the latest spec version, reconciling Excel inputs across people, redoing calcs after errors — 25% is usually closer to reality.
- Hourly rate. Use your team's fully-loaded rate (salary plus benefits plus overhead), not just salary.
The numbers above are the floor, not the ceiling
Every input is deliberately conservative — so your CFO can defend the number. The real upside is higher:
- Capture rate. Default 15%; engaged teams see 25–40%.
- Risk reduction. Modelled at one prevented event per year; weekly isn't unusual.
- Collaboration upside. Capped at 5%; materially higher at scale.
How this connects to a pilot
The ROI model is most useful when you can replace our default assumptions with your team's actual data. That happens during a pilot — see how a CalcTree pilot works for the structure.
Next step
If you'd like a custom version of this model with your team's numbers, we'll build it in CalcTree and share an interactive workspace you can flex the assumptions yourself.
