Carl Heinz Pierre on why occupancy is a lagging indicator

Occupancy is a report card, not a forecast. I am Carl Heinz Pierre, and I spent years running coworking space in Washington DC where occupancy was the number in every board deck, then moved into hospitality marketing where the equivalent number is RevPAR. Both describe a quarter that already closed. Neither one has ever told me what was about to happen, and the gap between those two jobs is where most marketing budgets get spent badly.

The number we reported was never the number that mattered

By 2015 I was responsible for three WeWork locations in Washington DC, roughly 1,500 people across them, work that Washingtonian covered when it named me to its list of the city's top tech leaders that year. Occupancy at those locations was strong. It was also the least useful number I had, because it moved after the decision rather than before it. A member who had privately decided to leave in January still counted as occupied until the contract ended in June. The metric was at its most reassuring exactly when it was least true.

The numbers that carried real information were smaller and less flattering. Renewal rate broken out by cohort, so a strong month of new signups could not hide a weak month of retention underneath it. How many members brought in someone they already knew, which is the cheapest growth there is and the hardest to manufacture. And how many people were physically in the building on a Friday afternoon, which was the closest thing we had to a live read on whether people wanted to be there or merely had somewhere to be. WAMU covered the DC coworking market in 2014 and framed the whole category as a question about how people wanted to work. That framing was correct, and it is a question occupancy is structurally incapable of answering.

Density is a product decision that keeps getting made as a real estate one

The obvious move when a floor performs well is to add desks. The math is clean, the incremental cost is close to zero, and the number you are graded on goes up immediately. What actually happens is that the product changes underneath you. Coworking sells proximity to people you want to be near, and past a certain density, proximity stops being the benefit and starts being the cost. Noise rises. The good tables are always taken. Phone rooms develop waiting lists. The members most worth keeping are the first to notice, because they are the ones with options.

That is the trap stated plainly: the change that improves the lagging indicator is frequently the same change that degrades the leading one. In my experience, curation is the product. Capacity is a byproduct of the product working.

What made this hard to catch in real time was that the damage never announced itself as damage. Nobody wrote in to say the floor had gotten too crowded. They renewed at a shorter term, or they stopped bringing guests, or they quietly took the meeting somewhere else. Every one of those signals existed in data we already had, and none of them appeared anywhere near the top of a report, which is a design failure rather than an analytics one.

Hospitality runs the identical trap with different vocabulary

Move to hotels and the words change while the structure stays exactly the same. RevPAR, ADR, and occupancy are the reported set, and all three look backward. They tell you what a quarter did. They do not tell you why it did that, and every one of them can be pushed in the right direction while the business gets weaker. Discount hard, sell through channels at a margin you would not accept if you examined it closely, and occupancy climbs on a chart while the underlying demand for the property stays flat or falls.

The leading set in hospitality has the same shape as the coworking one. Repeat guest rate, because a returning guest is the only person who has priced the experience accurately. Direct booking share, because it measures whether people are looking for the property or stumbling onto it. Referral volume, for the same reason Friday attendance mattered on a coworking floor. And the one I trust most, which is whether a guest who came for one reason comes back for a different one. A guest who books first for a conference and returns for a weekend has told you the place did the work, not the promotion. Those numbers move before revenue moves, which is the entire definition of a leading indicator.

None of this makes the reported set wrong. RevPAR is a real number and it belongs in the deck. The failure is one of hierarchy. When the backward-looking number sits at the top of the page, every conversation about the property starts from what already happened, and the team spends its energy explaining a quarter instead of changing the next one.

A test for telling one from the other

One question separates them. Can you move this number by spending money without changing anything about the actual experience? If the answer is yes, the number is lagging. Occupancy, RevPAR, site traffic, and impressions all fail that test, which is precisely why they are easy to hit and easy to fake.

Four things that make the distinction usable:

• Ask what the number looked like three months before the last thing that went wrong. If it looked healthy the whole way in, it is lagging and it will be lagging next time too.
• Separate volume from mix in every report. Growth that comes entirely from the cheapest channel is not growth, it is a trade, and trades should be approved rather than discovered.
• Pair them on the page. Report one leading indicator directly beside its lagging counterpart every month without exception, because the pairing is what makes a bad trade visible in a single glance.
• Give the leading indicator a named owner. Numbers without owners get explained rather than moved, and a metric nobody is accountable for is a metric that exists to be narrated.

The shift worth making

The shift worth making is to stop grading a place on how full it is and start grading it on whether people come back. Occupancy and RevPAR earned their place in the report. They have not earned the top of it, because by the time either one tells you something is wrong, the decision that made it wrong was made a quarter ago and the people who could have changed it have moved on to other work.

This runs well past hotels and coworking floors. Any business selling a place, a room, a seat, or a subscription has one comfortable number describing the past and one uncomfortable number predicting the future, and almost every reporting structure I have seen puts them in that order. I write more about measurement, hospitality, and the Washington region on carlpierre.com, including what coworking taught me about hospitality marketing.

Carl Pierre is a performance marketing strategist based in the Washington DC metro area. He ran three of the city's WeWork coworking spaces before moving into hospitality marketing, and his work has appeared in Washingtonian, on WAMU, and in Northern Virginia Magazine. More at carlpierre.com.