What 39 units taught me about underwriting maintenance

Every deal model I have ever seen prices the acquisition and forgets the decade after it. Cap rate, yield, service charges, maybe a sinking fund line. Then the deal closes and the real spending starts: an AC compressor here, a water leak there, a handover that needs 40 snag items closed before a tenant moves in. I run just under 40 residential units across Dubai, mostly Palm Jumeirah, Bluewaters and JVC, and the most useful thing that portfolio has done for me as an investor is hand me an accurate number for the line everyone else guesses at.

The number nobody models

Across our units, planned and reactive maintenance runs between 3% and 6% of gross annual rent, depending on the age of the building and who built it. Newer towers sit at the low end. Anything handed over before 2015 with original plant sits at the top, because chillers, pumps and waterproofing all start talking at once around the ten-year mark. That range is not a footnote. On a unit renting for 180,000 AED a year, the spread between 3% and 6% is 5,400 AED annually, every year, for the life of the hold. Underwrite the yield from 7% to 6.2% and the "great deal" is suddenly ordinary.

The first time this landed for me was not in a spreadsheet. It was a summer when three villas needed condenser replacements within one month, and the combined invoice was larger than the annual service charge on any single unit. Nothing exotic. Just old plant, Dubai heat, and no preventative program in the years before we owned them. When I went back through the acquisition files afterwards, the service records had been sitting there the whole time. Nobody had read them, including me.

Why I built a maintenance company instead of buying one

The standard answer is a facilities contract. We tried that route first and the problem was not price, it was accountability. Three vendors on one property means nobody owns the outcome. The AC contractor blames the leak on the plumber, the plumber blames the tiler, and the landlord pays twice. European Technical started as an internal fix for exactly this: one number to call, one party responsible, across AC, plumbing, electrical and handyman work, originally around our own units on the Palm and at Bluewaters. It now runs as a commercial business serving other landlords across Dubai and Sharjah, which was never the plan, but the demand told us the problem was universal.

The investor's takeaway is not "build a maintenance company." It is that whoever holds the maintenance accountability holds a big share of the real return. If you outsource it, your underwriting is only as good as your vendor's honesty about what is coming.

Do not confuse maintenance with service charges. The service charge, which on Palm apartments currently runs between 18 and 30 AED per square foot per year, pays the building for shared plant, security and common areas. It does nothing for the compressor inside your ceiling or the pump under your kitchen sink. I have watched buyers subtract service charges from rent, conclude maintenance is "included", and discover in month seven that in-unit plant is their problem alone. The two costs stack. Model them separately or you are quietly overstating yield by a full percentage point, which across a 39-unit portfolio is the difference between a good decade and a mediocre one.

Three things I check before buying now

First, the age of the plant, unit by unit. Ask for the last two years of service records, not the brochure. A chiller installed in 2012 with no compressor replacement on record is a liability with a delivery date, and the record tells you before the surveyor does.

Second, water history. In our portfolio, water incidents, burst flexi hoses, shower tray failures, AC condensate blockages, have caused more aggregate damage than every electrical fault combined. A 60 AED flexi hose can take out a 40,000 AED floor. Check whether the unit has isolation valves that actually turn, and ask when the hoses were last changed, because almost nobody changes them on schedule. We now replace them at every turnover as policy, and it is the cheapest insurance in the building.

Third, the tenant handover pipeline. Vacancy cost is obvious, but the snag-and-repair window between tenants is where hidden budget goes. We budget 8,000 to 15,000 AED per unit turnover for a mid-market apartment, more for a furnished one, and we are rarely under. If your model shows turnover costing nothing, your model is fiction.

The boring conclusion

None of this is sophisticated. It is a percentage, a parts schedule and a habit of reading service records. But the returns in residential property are rarely won at the negotiation table. They are won in the years of small decisions after it, and most of those decisions are maintenance decisions. Underwrite them honestly, hold someone accountable for them, and the asset performs roughly as modelled. Ignore them and no purchase discount will save the IRR.

One last habit worth copying: review the actual spend against the underwritten percentage every quarter, per building, not per portfolio. Averages hide everything. Two of our towers run under 3% and look after themselves; one JVC block runs near 7% because of a builder who valued finishing over waterproofing, and it is permanently on the watch list. The quarterly review is 20 minutes in a spreadsheet and it is the only reason I can quote that range with a straight face.

I keep notes on the operational side of these businesses in the journal, including the five numbers I watch weekly in the service business that grew out of this problem.