Unit 5 · Maximizing Performance Through Iteration · Lesson 5.11
When you are booking up too fast
A calendar that is full three months out is not a win, it is a pricing signal. This two-bedroom in downtown San Antonio took 117 nights of bookings in one two-week window, most of them far out at around $90 a night. The diagnosis is straightforward, the correction is a lift at every step, and the proof takes months because the calendar is already sold.
The full lesson text below is an edited transcript of the video, published 2026-08-24. The complete course is free at the playbook.
The listing and the symptom
A two-bedroom property in downtown San Antonio, Texas. Its steps split the booking window in the standard way: a hedge covering zero to one days out, a control covering two to 26 days, and a test covering 27 to 365 days. Cleaning runs at $100 a stay, tax at 13.5%, and no service component is added on top. Those numbers are recorded before anything else, so the rates the listing actually achieved are compared against market prices on the same basis.
Conversion is healthy. Search to listing conversion is good and listing to booking conversion is high, in fact higher than it ought to be, which is the first hint that something is wrong. That check always comes first. If a listing is not converting, the problem is the listing, and no amount of rate work will fix it. Here the listing converts beautifully, which points the investigation straight at price.
The owners had built a pricing strategy and implemented it. They were very conservative in setting their three steps, and the result was a slew of bookings. The job now is to bring the strategy back into alignment with what actually happened, so that nothing further is left on the table.
Where the bookings landed
The first read of any iteration is simply where in the booking window the bookings arrived. Across the two-week period ending 25 November there were four bookings in the hedge, seven in the control and 19 in the test. Thirty bookings, 117 nights on the calendar.
The shape matters far more than the total. A few last minute bookings in the hedge is normal. A solid block through the control is exactly what a working strategy produces. Nineteen of 30 bookings landing 27 or more days out is not. Guests booking that far ahead are the least price sensitive segment there is, and if they are filling a calendar at test prices, the test prices are too low.
That skew also tells you the underpricing is unlikely to be confined to the test window. Any peak dates sitting inside it will have gone out at the same discount, sold months early and locked in at the wrong number.
Reading occupancy against the market
The second read compares occupancy and price position, step by step, against what the market is doing over the same dates. In the hedge the occupancy index is 3.19, more than three times the market. Through the control it runs at roughly four and a half times. Through the test it sits between three times and nearly five times.
An occupancy index measures against the market, not against a full calendar, so the raw percentages can look modest and still be alarming. Test occupancy is only 13% midweek and 21% on weekends, but that far out the market has barely sold anything, so those numbers are three to five times the market rate. A high index in the hedge on its own is fine, it just means you are booked. Three high readings across all three steps is the diagnosis.
One reading is worth noting and then setting aside. Occupancy across the 90 day window is running about 10% below the same period a year earlier, so the market itself is softening and pacing behind. That is not this property’s problem. This property is being booked at cheap prices, and the correction is upward regardless of which way the market is drifting.
The hedge, midweek only
A hedge that spans zero to one days is two days wide, so unless a weekend happens to fall inside it, only midweek dates appear. Here none do, which leaves the weekend hedge price unanswered until later in the process.
For midweek the picture is clear enough. Expected market occupancy over those dates is about 35%, the listing has been getting booked at about $102 a night, and its occupancy index is 3.19. Booked more than three times as heavily as the market, at a price sitting below the middle of it. The midweek hedge moves to $110, closer to the 50th percentile.
The control, effectively sold out
Control occupancy is 78% midweek and 100% on weekends. Measured against the target range for a control step, that is far too high, and the weekends are the worst of it. Over the same dates the market still has about 15 points midweek and 25 points on weekends to travel between the occupancy it holds now and the occupancy it is expected to finish at, so the market has a long way left to run while this listing has almost nothing left to sell.
Midweek rates sit at about $90, between the 25th and 50th percentile of the market. With occupancy near 80% that is plainly too cheap, so midweek control goes to $120, above the 50th percentile. Weekends have been booking at $106 at 100% occupancy, with the market expected to reach about 60%.
The first instinct on weekends was $140. The number actually set was $145, and the reasoning is worth copying. When a calendar is this full, the only nights left to sell are the ones a cancellation hands back, and those few nights should be waiting at the highest price the market will support.
The test window, where it leaked
Test rates were $90 midweek and $112 on weekends. That is what pulled 19 bookings in from 27 days and further out, and it is where most of the lost revenue sits.
Midweek, market occupancy over these dates reached about 50% in the same period a year earlier, which is significant demand to be selling into at $90. The midweek test price goes to $135, around the 50th percentile.
Weekends a year earlier ran between roughly 55% and 69% occupancy. The weekend test price has to sit above the $145 weekend control, since nights further out should not be cheaper than nights closer in, so add about 15%, a little over $20 a night, and round it to $165. That lands above the 50th percentile without being far above it.
Answering the weekend hedge
The weekend hedge price still has no data behind it, so the month’s day by day numbers fill the gap. The dates in question fall in November, and Thanksgiving has to be excluded first. Occupancy through a holiday peak is not representative of an ordinary weekend, and including it would drag the hedge price somewhere it does not belong.
Across the remaining weekends, the 25th percentile averages about $98.50 and the 50th percentile about $125. Occupancy on those same weekends runs between 56% and 60%, which is decent for a quiet stretch of the year and argues for pricing nearer the middle of the market than the bottom of it. The weekend hedge is set at $125.
One check before committing to it. A hedge price must not exceed the control price for the same day type, or the ladder inverts and the nights closest in become the dearest on the calendar. At $125 against a $145 weekend control, the ladder still rises correctly.
The new price ladder
A fresh iteration opens on 26 November and runs two weeks to 10 December, with the step lengths unchanged. The hedge is $110 midweek and $125 on weekends. The control is $120 midweek and $145 on weekends. The test is $135 midweek and $165 on weekends.
Against what the listing had been achieving, that is a lift everywhere: roughly 8% on the midweek hedge, a third on the midweek control, and close to half on both test prices. It is a significant increase across the board, and it is justified purely by the volume and the position of the bookings that came in.
The individual moves look subtle because the market itself is compressed. The 25th percentile is $88 and the 90th percentile is $182, so there is not much distance between the cheap end and the expensive end. This is a low end market, and the game is not to chase a premium it will not pay. It is to stay heavily booked at a smarter number than the one you were using.
Booked, at the wrong price
This is the trade between occupancy and average daily rate in its plainest form. The listing was never in trouble. It was booked, the owners could see bookings landing, and everything looked like it was working. It was working at the wrong price.
The next iteration will look flat
There is a consequence to the diagnosis that catches people out. Because control occupancy is already sitting between 78% and 100%, the next iteration will produce very few bookings. The nights are gone. Only a couple of midweek control nights remain unsold, and the hope is that those go at $120.
Run the iteration anyway. The discipline does not pause because the data is thin. You still look at where the bookings came in, still compare occupancy against what the market is expected to achieve, and still reset the rates against comp set conditions at every pass. An anemic iteration is a reading, not a reason to stop reading.
Then accept the lag. There is no way to tell whether $145 weekends and $165 test nights are working until the calendar frees up, and for the time being it simply is not available. A booked night is booked, and short of a cancellation there is no opportunity to change what it earns. Nothing will move overnight, because the inventory was already sold at the old prices.
So hang tight, and stay on top of it. Being overbooked cheap is an uncomfortable position but a fixable one. The only way to make it worse is to overcorrect out of impatience and end up staring at a completely vacant calendar instead.