Corporate Real Estate Portfolio Software Problems: The 7 That Cost Real Money, and How to Avoid Them

Why does the lease register swallow the whole project?

Because the lease is the document, so it feels like the object. Every scoping conversation starts with the abstract: which fields, which clause types, how many lease forms, how do we handle the amendments. Six weeks later the data model is a lease with a landlord and a property attached, and the project has quietly become a better version of the register you already have.

An occupier does not make lease decisions. It makes location decisions. Renew, relocate, downsize or exit is a question about a place, and answering it needs the lease, the trading or utilisation performance, the capital already sunk, the remaining useful life of the fit out, the market alternatives, the restoration obligation and the cost of being wrong. The lease is one input among six. When the primary object is wrong, every report is subtly answering a question nobody asked, and the report that finally gets requested is built in a spreadsheet by an analyst anyway.

The fix is to make the location the primary object at the data model stage and hang everything from it, including the lease. That single decision changes what the system can do later. It is also the decision that is nearly impossible to reverse without rebuilding, which is why it belongs in the first two weeks rather than in a phase two conversation. If a scoping document you are reading describes leases with locations attached, send it back before anyone writes code.

What goes wrong with lease abstraction and the data you migrate?

Abstraction is the pacing item on almost every portfolio project above a thousand leases, and it is routinely priced as a data load.

The specific problems are consistent. Abstracts prepared for an accounting purpose capture what accounting needed: commencement, term, rent steps, options as a date. They do not capture the notice mechanics, the conditions attached to exercising an option, the recovery caps and exclusions, the restoration standard, or the co-tenancy trigger, because none of those produce a journal entry. Migrate that abstract into a decision system and you have imported the fields that were never the problem while leaving the fields that cause the losses in a PDF.

Second, abstracts age badly. Amendments, side letters, consents and estoppels change the position and frequently never reach the abstract, so the system is confidently wrong. A missed side letter that varies a notice period from nine months to twelve is exactly the kind of detail that turns a well run process into a missed option.

Treat abstraction as a funded phase with a quality standard, not a task. Use an existing professional abstraction where you have one rather than re abstracting everything, since that is the most reliable way to shorten the timeline, but audit a stratified sample against the source documents and measure the error rate before you trust it. Capture the source document reference and the abstraction date on every field so a user can see how old the underlying reading is. And build a mechanism for amendments to update the abstract as part of the transaction workflow, otherwise the decay starts again the day you go live.

Why do the finance, sales and construction feeds break after launch?

Because they are built as one time imports against a sample file and then meet a production reality that changes.

The finance feed carries the payments and the charge history, and it breaks on the join. Your finance system identifies a site by a cost centre, your property system by a lease identifier, your point of sale (POS) or branch reporting by a store number, and none of them agree. Somewhere there is a mapping table maintained by a person. When a site is refurbished and reopens with a new store number, or a legal entity is restructured, the mapping silently stops matching and the occupancy cost ratio for that site goes wrong without anyone being told.

The trading feed breaks on definitions. Sales net of what, over which calendar, including or excluding transfers and online orders fulfilled from the store. Two teams can both be right and produce different numbers, and the portfolio system inherits whichever one it was connected to.

The construction feed usually breaks because it was never a system. It is a spreadsheet maintained by a project manager, and the integration was a promise to export it monthly.

What works: an explicit site identity service that owns the mapping between every external identifier and your location record, with unmapped records surfacing as an exception queue rather than disappearing. Agreed definitions written down and stored with the feed, so a change in the source definition is a visible event. And a reconciliation job that reports feeds which have gone quiet, because the failure mode you cannot see is a feed that stopped delivering and left the last known figures on screen looking current.

What happens when restoration and co-tenancy obligations are not covered?

They surface at the worst possible moment, which is at exit, priced by the counterparty.

Restoration and make good clauses require you to return the premises to a defined condition. If that liability is not estimated per site and carried forward, then every exit decision the business makes is understated by the cost of the exit, and the model recommends closing sites it should keep and keeps sites it should close. Discovering the number during a dilapidations negotiation is discovering it from the landlord's surveyor.

Co-tenancy rights are the mirror image. An anchor leaves a centre, a right arises, and the right expires if not exercised within a window. A system that holds the clause as text in an abstract has recorded the right without creating any prospect of it being used.

The rest of this category is administrative and therefore neglected: percentage rent reporting on landlord specific schedules, consent processes with response deadlines, continuous operation clauses a temporary closure can breach, and annual insurance certificate delivery. Individually each is small, which is why a stretched team drops them.

Model obligations as recurring or triggered items with an owner and required evidence rather than as notes. Carry an estimated restoration liability per site, refreshed on a cycle, so it appears inside exit economics automatically. Automate certificate delivery entirely, because it is pure administration and should never consume a person.

Should you build custom or configure what you already own?

Configure, and we will say it plainly, if you occupy fewer than about 150 locations or if your driving requirement is lease accounting compliance. CoStar Real Estate Manager and Accruent Lucernex do lease accounting properly, come with auditor familiarity, and cost far less than building that capability from nothing. Building your own accounting engine to save a licence fee is a poor trade.

Configure if you are a large owner occupier whose real problem is facilities, space planning and work orders across corporate buildings. That is what IBM TRIRIGA exists for, and rebuilding an integrated workplace management system is a large use of capital for a problem somebody already solved.

Look hard at Tango before building if you are a retailer, because it was designed around the site decision workflow described in this guide and it may fit closely enough that the gap is a reporting layer rather than a platform.

Build when the constraint is decision throughput rather than record keeping: when you have missed option or kick out windows and the cost was material, when each site decision takes an analyst a week so most of them are not made properly, when your landlord charges are paid without validation, or when your decisions depend on modelling transfer of revenue between locations, which no packaged occupier tool does with your data.

How do hidden costs get into the quote?

Through five doors, in rough order of size.

• Country count. Lease structures, indexation, tax treatment and statutory renewal rights differ enough between jurisdictions that each region is genuine scope, not a configuration flag. A proposal priced for one country and delivered across four will be renegotiated.
• Lease accounting. If it is in scope it brings remeasurement on modification, discount rate policy and auditor evidence with it. That is a work package, not a report. Decide explicitly at the start, because retrofitting it later costs more than planning for it.
• Abstraction backlog. Above a thousand leases this is usually the pacing item and it is frequently absent from the engineering quote entirely.
• Integration count. The system is only as good as its feeds, and each one carries identity mapping, definition agreement and ongoing reconciliation rather than a single connector.
• Property type mix. Offices bring space and occupancy planning. Stores bring trading analysis. They share almost no logic, so a portfolio with both is closer to two builds than one.

Ask any bidder to price these as named lines against your actual portfolio, and treat a proposal that omits all five as incomplete rather than competitive.

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

Backward scheduling. The system must work backwards from each notice deadline through your real decision durations, taken from your own history, and open a case at the point the process must start. If a renewal needs a market test, a capital assessment and a committee, the case opens twelve to eighteen months out, not on the date.

A populated case. The accountable manager should open a decision, not a research task. Trailing sales or utilisation, occupancy cost against your portfolio benchmark, remaining fit out life, capital spent in the last five years, comparable market rents and the restoration liability if you exit, all present before they arrive.

A named owner and an escalation path, with sites where no decision is recorded by the point of no return reported as a board visible number rather than an internal embarrassment.

Landlord charge validation built in the first phase rather than deferred. Encoding recovery caps, exclusions and share definitions once at abstraction, then testing every incoming statement against them and ranking a review queue by exposure, is the feature your finance director understands immediately and the one most likely to pay for a meaningful part of the build within two reconciliation cycles.

And ownership settled in writing before kickoff: the repository, the infrastructure accounts and the unrestricted right to hire another firm. A portfolio system holds commitments running decades into the future and should not depend on a vendor relationship with a shorter expected life than the leases it manages.