Facility Condition Assessment Software Problems: The 5 That Cost Real Money, and How to Avoid Them

Why does the assessment get scoped as a line item list?

Because that is the shape of the deliverable. Eleven volumes, a spreadsheet appendix with 26,000 rows, and a headline backlog figure. A developer given that spreadsheet builds a database with a login screen around it, and every structural weakness of the consultant deliverable is preserved with better styling.

The permanent record has to be the asset, not the finding. Each building needs a hierarchy down to system and component level, and every assessment produces observations against those components rather than free standing rows. Then a chiller has a history: installed 2004, assessed fair in 2019, repaired twice, assessed poor in 2024, replaced in 2026. The Facility Condition Index derives from that history rather than being restated from scratch every cycle, and the follow up survey becomes an update rather than a fresh start.

Ask a prospective developer to draw the model before you sign. You should see building, system and component with an identity that persists across assessment cycles, plus observation, deficiency, renewal event and funding source as separate objects. If the sketch is buildings and repairs, you are paying to rebuild the consultant spreadsheet. The same conversation should surface the things that never fit a product: leased space you maintain but do not own, assets on a ground lease, historic structures with restricted intervention, and campus infrastructure such as steam tunnels and distribution that sits inside no single building.

What goes wrong when you merge several consultant assessments into one history?

This is the item most often underestimated, and it is not a technical problem. Two consultants surveyed your estate five years apart using different classification depths, different unit cost sources and different naming, so the same rooftop unit appears as one line in 2019 and three in 2024, or as a system level deficiency in one and a component level one in the other. Matching them is judgement work performed by somebody who knows the buildings, not a join on a column.

Do it selectively. Map the assets that carry your money, typically roofing, primary heating and cooling plant, electrical distribution and elevators, and accept that low value line items may not reconcile. A partial history that is trustworthy beats a complete one that is invented, and an invented match produces a service life comparison that will mislead you for a decade.

The second trap is forcing your estate into a standard hierarchy. Map to a recognised classification for external reporting and grant applications, because you will need it, but carry your own asset classes as first class objects underneath. A health system with central utility plants, medical gas and a linear accelerator vault does not fit the same tree as a university with residence halls and a stadium. When classes do not exist, assessors put them in Other, and three years later Other is a fifth of your backlog and useless for planning. Institutions that already hold a clean space and asset inventory move through this phase considerably faster than those starting from spreadsheets.

Why do maintenance system and mapping integrations break after launch?

The read half is straightforward. Maximo, AiM, TMA and Archibus all expose work order history, and pulling it works. The half that breaks is classification: deciding which work counts as renewal against which counts as routine maintenance. That is a judgement your organisation has to make explicitly, and if it is left implied, the rule drifts. A shop starts coding compressor replacements differently, a new work type appears after a reorganisation, and renewals quietly stop being recognised. Your backlog then stops retiring even though work is being done, which is precisely the appearance you are trying to avoid in front of a board.

Make the classification rule visible and owned. Publish it, review it annually with the maintenance managers, and report the volume of work orders that could not be classified rather than dropping them. A rising unclassified count is an early warning that your bridge is about to stop reconciling.

Writing renewal completions back is the harder direction, and it is where estimates are optimistic. If the maintenance system and the capital plan are to agree, a completed renewal has to close the component's deficiency and update its installed date, and that write path needs to handle partial completions and cancelled projects. Geographic information system layers for campus infrastructure are a fifth integration with their own model, not an extension of the fourth. Ask which specific system and which version a developer has worked with, because the data models genuinely differ.

What happens when funding restrictions and the backlog bridge are not covered?

A single prioritised list is not executable in a public institution. Bond proceeds, state appropriations, grant funds, auxiliary reserves and departmental budgets carry different restrictions on what they may be spent against, so the top item on your list may be ineligible for the only money available. Attach eligibility rules to funding sources, generate a plan per source, and then show what remains unfunded across all of them. Without that, your capital plan is a wish list that your finance office quietly reorders.

The bridge is the other omission and it is the one that costs credibility. A trustee sees significant spend and an unchanged or higher backlog, and both facts are usually true, because the spend went to items outside the assessed backlog, or to renewals never written back, or the backlog grew through escalation and new findings faster than it was retired. Without a reconciliation you look either incompetent or evasive.

Build the bridge as a first class output: opening backlog, plus escalation, plus new findings, minus completed renewals, minus items reclassified or removed, equals closing backlog. Generated live from linked work orders and capital projects, it answers the question in the meeting rather than in a fortnight. If a developer cannot describe that bridge when you ask, they have not sat in the meeting where it was demanded, and you will be the one holding the silence.

Should you build custom or configure what you already own?

Stay with a product if you are under roughly 1 million square feet with a stable estate, a single funding source and no legislative or trustee reporting obligation beyond an annual figure. Brightly Predictor is a reasonable fit there, and commissioning a Gordian assessment on a cycle is a perfectly defensible way to run a small portfolio. Custom software would be an expensive way to hold 26,000 rows.

Before commissioning anything, look hard at what you already licence. Accruent VFA and Gordian both bring real methodology and strong cost libraries, Gordian in particular through the unit cost data behind it, and a meaningful share of what gets reported as a limitation is an unmaintained implementation: an asset hierarchy set up once and never revised, scenarios nobody was trained to run, and reports that stopped working when a key analyst left. Ask whether your complaint is that the tool cannot express your estate, or that nobody has updated it since the last assessment.

Build when two or more hold. Your estate is large enough that the assessment cycle costs more than a small team. Your asset classes genuinely do not fit a product hierarchy, which is normal for health systems, ports, transit agencies and research campuses. You must reconcile backlog against completed work in front of a board or a legislature. You hold multiple funding sources with different restrictions. Or your renewal decisions have started diverging from the assessment because everyone knows it is out of date, which is the clearest signal the document has stopped being a system.

How do hidden costs get into the quote?

Asset class count is the first. A quote priced against an office portfolio and delivered against clinical infrastructure, central plants and campus distribution is a different model, because each class brings its own condition criteria, service life behaviour and renewal logic.

Maintenance system integration depth is the second, and it is usually quoted as one line covering the easy direction. Reading work orders is fast. Writing renewal completions back, with partial completions and cancelled projects handled, is the part that adds weeks.

Offline field capture is the third. If assessors work in basements, tunnels and remote sites with no signal, the inspection app has to complete, photograph and sign work with no connectivity and synchronise cleanly afterwards. That is real engineering rather than a checkbox, and it is often assumed to be free. Fourth is cost library licensing, if you want commercial unit cost data rather than your own historical bid data, which is a recurring line rather than a one off. Fifth is the multi assessment migration, which is your staff's judgement time as much as the developer's. Sixth is board and legislative reporting formats, which arrive late in the project as an assumed deliverable and are rarely as simple as the three charts everyone imagines.

What separates a build that works from one that fails here?

The ones that work start with the top fifty buildings by replacement value and the three highest risk systems, typically roofing, primary heating and cooling, and electrical distribution. That covers most of the money, teaches the model on data you understand, and produces a defensible number in one budget cycle. The ones that fail attempt the full estate, the field app, maintenance reconciliation and scenario modelling together, and present from a spreadsheet in the meantime.

Use published service life as a default and then beat it with your own evidence. Allow per asset adjustment with a recorded reason, and after a few years compare actual replacement ages against assumed life by asset class and exposure. Learning that your 1970s residence hall roofs last one span and your 1990s academic roofs last considerably longer is the most persuasive evidence available in a funding hearing, and no purchased table can tell you that.

Make scenarios first class objects rather than reports. Flat funding, a bond issuance, targeted reinvestment in the worst quartile, demolish and consolidate, each projecting backlog and index by building and by portfolio on your fiscal calendar. Add project bundling, because forty roof line items across a campus are one procurement, and a plan that cannot bundle produces a list nobody can execute.

Settle ownership of the code and the asset data in writing before kickoff, including the repository, the infrastructure accounts and the right to appoint another firm. At Digital Heroes the client owns both from the first commit. The multi decade asset history you accumulate is an institutional record that will outlive any vendor relationship, and it should never sit behind somebody else's licence.