v8.8.0 - Monte-Carlo cost-risk for a bill of quantities, and an in-app How it works hub.

An estimate that resolves to one figure tells you what you expect to spend, but not how confident you should be in it. 8.8.0 answers the second question. Its headline is a Monte-Carlo cost-risk analysis that turns a bill of quantities into a distribution rather than a point, so you can talk about a confidence level and a contingency instead of a single guess. Alongside it, a How it works hub puts a plain-language guide to every module inside the app, and a clutch of smaller additions and fixes round out the release.

What is new in version 8.8.0

• Monte-Carlo cost-risk for a bill of quantities. It runs thousands of correlated, PERT-distributed iterations to produce a full cost distribution rather than a single point estimate: P5 to P95 percentile bands, mean and standard deviation, a probability S-curve (the chance the total lands at or under any given figure), a recommended contingency at your target confidence level, and a tornado chart showing which positions drive the most variance. A one-factor correlation keeps systemic risk from cancelling out across lines.
• An in-app How it works hub. Reachable from the Help menu, it explains every module - what it does, the main steps to use it, and a few practical tips - translated into all 27 languages.
• Lightweight 2D maps in the Geo Hub. A basemap switcher lets you place and review project locations without loading the full 3D globe.
• A DIN 276 element breakdown in cost benchmarks. With a short, plain-language guide to reading it.
• Count by example in takeoff. Pick one symbol on a drawing and the tool finds and counts the matching symbols across the sheet.

An estimate as a distribution, not a single number

Every line in a bill of quantities carries uncertainty, and those uncertainties do not simply average out - some move together. The new Monte-Carlo cost-risk analysis takes that seriously. It runs thousands of correlated, PERT-distributed iterations across the bill and produces a full cost distribution: the P5 to P95 percentile bands that frame the likely range, the mean and standard deviation, and a probability S-curve that reads off the chance the total lands at or under any figure you point at.

From there it answers the question a client actually asks - how much contingency - by recommending one at your target confidence level, and it shows a tornado chart of which positions drive the most variance, so you know where to tighten the estimate first. A one-factor correlation keeps systemic risk from cancelling out across lines, which is the failure mode of a naive simulation that treats every line as independent. The result is a risk picture you can defend, not a single number with a gut-feel margin bolted on.

The whole platform, explained from inside it

A platform with this many modules needs a way to learn it that does not mean leaving it. The new How it works hub, reachable from the Help menu, explains every module in plain language - what it does, the main steps to use it, and a few practical tips - and it is translated into all 27 languages, so the guidance reads correctly whichever language a team works in. It turns the app into something you can explore with the manual open in the same window.

Three smaller additions round out the release. The Geo Hub gains lightweight 2D maps with a basemap switcher, so you can place and review project locations without spinning up the full 3D globe. Cost benchmarks gain a DIN 276 element breakdown with a short guide to reading it. And takeoff can count by example: pick one symbol on a drawing and the tool finds and counts the matching symbols across the sheet.

Fixes and hardening

A bill-of-quantities parent position now rolls its children's progress up as a quantity-weighted average - falling back to a simple average when the children carry no quantity - so a parent's percent-complete reflects the relative size of its parts instead of treating every child equally. Contract cumulative completed value is recomputed on the server, so progress claims always reconcile to the stored line items, and a quality pass across the lower-traffic modules fixed currency display, action-button gating and several save and persistence issues.

Installing a BIM or CAD converter no longer fails with "signal timed out" on a slow server: the download now runs in the background and the panel updates when it finishes, instead of the request being cancelled mid-download. On the security side, the ERP chat assistant and the project-intelligence advisor now honour project team membership when checking access, the property-development broker leaderboard scopes strictly to the caller's own brokers, and the bundled build-time dependencies were updated to clear all known advisories - twelve in total, five high-severity, none of which ship in the running application.

By the numbers

Install or upgrade

pip install --upgrade openconstructionerp

The desktop installers for Windows, macOS and Linux carry the same one-installer setup, and the Linux build includes the CAD and BIM converters for AutoCAD, Revit and IFC. You can grab the latest installers from openconstructionerp.com/download. If you run an external PostgreSQL through DATABASE_URL, nothing about that connection changes. Questions or trouble upgrading, write to info@datadrivenconstruction.io.