Build Flows

Forecasting and insights

Construction forecasts that show their assumptions

We build forecasts from the data your project managers and accountants already keep: the cost-to-complete forecast, issued invoices and bills, contracts and signed work. Each projection states the rules behind it, month-end snapshots keep reported figures fixed so trends are honest, and jobs that drift are flagged for review. A forecast is only as good as its inputs and assumptions, so we make both visible.

The value it creates

  • Forecasting

    Project forward with stated assumptions

    Cash, backlog and margin projections are calculated from written rules, such as which invoices count as expected collections, so anyone can see why a number is what it is.

  • Early warning

    See fade before month-end close

    Jobs with no forecast, a negative margin, or cost above the estimate at completion are flagged on every refresh, so a project manager updates the forecast before the WIP is published.

  • Better decisions

    Decide on numbers that hold still

    Month-end snapshots keep reported figures as reported, so a board pack from March still matches March and the change since then is visible on its own.

  • Visibility

    From portfolio to cost code

    Start with backlog, margin and cash across all jobs, then drill into one project's budget, committed cost, cost to date and estimate at completion by cost code.

  • Time saved

    Less time rebuilding the forecast workbook

    The forecast refreshes from source systems on a schedule, instead of someone re-keying exports into a spreadsheet each month.

  • Standardization

    One definition of every forecast measure

    Percent complete, backlog months, fade and risk labels are defined once with the controller and calculated the same way for every job.

What we build

  • Cash forecast

    Expected collections and payments by week from open invoices and bills, with the cumulative cash position and the items behind each figure.

  • Backlog and burn

    Signed work not yet earned, by project and by month, with backlog months showing how long current work lasts at the recent earning pace.

  • Fade and gain tracking

    Month-on-month change in gross profit percent at completion for each job, calculated from stored month-end snapshots.

  • Early-warning flags

    Rule-based risk labels such as no forecast, at risk and watch, plus exceptions for jobs over their estimate at completion, so attention goes where it is needed.

  • Month-end snapshots

    One stored row per project per month, so trends, fade and comparisons use figures as they were reported rather than as they look today.

  • Scenario assumptions

    Assumptions such as collection timing or burn rate kept in a governed table your team can review and adjust, with the forecast showing which assumption set it used.

  • Forward work and capacity

    Backlog shown next to weighted CRM pipeline, never added together, alongside labor hours and utilization, so you can see whether upcoming work fits your crews.

How it works

Forecast: history and current commitments feed explicit assumptions about receipt and payment timing and burn rate, which drive several scenarios, producing a forecast with a range that is reviewed and snapshotted for later comparison to actuals.
  1. 1.History + commitments

    Actuals to date plus what's already contracted.

  2. 2.Assumptions

    Receipt and payment timing, burn rate — written down, editable.

  3. 3.Scenarios

    Base, slow-pay and fast-burn cases side by side.

  4. 4.Forecast with range

    A likely figure with a low–high band, not a single number.

  5. 5.Review & snapshot

    Signed off, then frozen so you can compare to actuals later.

  1. 1

    Agree the definitions

    We write down with your controller how each forecast measure is calculated, including edge cases like a job with no estimate at completion or cost above it.

  2. 2

    Connect the inputs

    We pull the project forecast, contracts, invoices, bills and pipeline from source systems into a lakehouse and join them on a project crosswalk.

  3. 3

    Validate before every refresh

    Data-quality checks run first; missing forecasts stay blank rather than becoming zero, and unmatched projects appear on an exceptions page.

  4. 4

    Snapshot month-end

    Each close stores the reported figures, so fade, gain and trend lines compare like with like.

  5. 5

    Publish forecasts and flags

    Power BI pages show the cash forecast, backlog and burn, fade and gain, and risk labels, each with the assumptions and sources it used.

Systems we work with

  • Procore
  • QuickBooks Online
  • Sage 100 Contractor
  • HubSpot
  • Outbuild
  • Primavera P6
  • Microsoft Fabric
  • Power BI

How we measure the value

We agree a baseline before we build and measure the same things after go-live. Use the monthly report cost calculator to put your own numbers on it.

What we measureHow baseline and after are captured
Forecast accuracy over timeWe store each month's projection alongside the month-end snapshot, then compare forecast cash and margin against what actually happened in later months.
Jobs with a missing or stale forecastThe refresh counts jobs with no estimate at completion or an outdated forecast, starting from the first run, and tracks the count each month.
Lead time on margin fadeWe record when fade on recent jobs was first raised under the current process, then compare with when the fade flag first appears in the report after go-live.
Hours to produce the forecastWe log the time your team spends building the cash and WIP forecast now, then track the same work once it refreshes from the pipeline.
Estimate your monthly report cost

Use cases

Frequently asked questions

Is this predictive analytics or AI forecasting?

Mostly it is not, and we say so. The forecasts are calculated from your own data and written rules, such as the project manager's estimate at completion and the due dates on open invoices and bills. That makes them explainable and checkable. Where a statistical or AI model would genuinely help, we discuss it during discovery and test it against your history before relying on it.

How does the cash forecast work?

In our production build the cash forecast uses committed cash only: expected collections from issued invoices and expected payments from bills in accounting, shown by week with the cumulative position. Pipeline deals and unbilled work are not counted as cash. Other assumptions, such as billing timing on unbilled work, can be added as stated, adjustable inputs.

Why do month-end snapshots matter?

If a report only holds the current forecast, past months change every time an estimate moves, and your earlier board pack no longer matches. Storing one row per project per month keeps reported figures as reported. Fade and gain are calculated from those snapshots, so the trend is honest.

What counts as an early warning?

Rules you agree with us, for example a job with no forecast entered, gross profit at completion below zero or below a watch threshold, or cost to date above the estimate at completion. Each flag says why it fired. The thresholds are yours to set and change.

Can we include CRM pipeline in the forecast?

Yes, but it is kept separate. Signed work belongs in backlog; a deal at a given probability does not. We show backlog and weighted pipeline side by side as total forward work, never added together.

How is the project priced, and how do we start?

Scope and price are fixed, agreed after discovery. A sensible start is your project management and accounting systems, the WIP schedule, an exceptions page and month-end snapshots; cash forecast, backlog and capacity follow once the core numbers are trusted. The WIP and Cash Forecast offer is built around this.

Next step

Which report or workflow would you like to improve?

Tell us what your team does today, which systems are involved, and what you want to change. We'll discuss whether there is a practical fit.

Prefer email? charley@buildflows.ai