How should I set my prices? Are they too low? Too high? What should my cleaning fee be? How much higher should I set weekends? What should I do in the low season? How much can I charge on holiday weekends? Can I raise the price when there are special events?
These are incredibly difficult topics even if you just have one property on Airbnb. It just gets harder as you add more: you still don’t know the answers, and the work just multiplies.
Fortunately there are tools that can help you with pricing, both by recommending what prices to set and in automatically pushing those prices to Airbnb. And Airbnb offers SmartPricing. In fact, Airbnb pushes SmartPricing pretty hard. Why not just use it?
Why Not Use Airbnb’s SmartPricing?
We’re now entering a scary new age where computers can actually be smart. Since I actually have some experience in artificial intelligence and machine learning, I know what it takes to succeed:
- Lots of data
- The smartest PhD data scientists, and lots of them
- Lots of computing power
In other words: boat-loads of money and the “cool factor” required to attract the very best. It goes without saying that Airbnb wins on every point.
And yet, many people, including myself, have tried Airbnb’s SmartPricing and have found the prices to be way off.
The Alternatives to Airbnb’s SmartPricing
The independent tools are Wheelhouse, PriceLabs, and BeyondPricing. All three work about the same way:
- You set a “base price”, which is the average of the daily prices that you want to set across the whole year.
- They predict how demand will fluctuate day-by-day over the next year.
- They use the demand prediction combined with secret magical algorithms to determine the delta, plus or minus, to apply to your base price for each day.
- They automatically update Airbnb with your prices.
Let’s take a specific example. Here’s one of my nicest 1BR/1BA apartments in Fort Lauderdale:
Let’s say I set a base price of $100/night. Why $100? Each of the services helps you with that, and we’ll get to that in detail later.
BeyondPricing for a $100 Base Price
So based on the $100 base price, here’s what BeyondPricing recommends. As you can see, it’s wildly different than a flat $100/night. The peaks are weekends, holidays, and events in the Fort Lauderdale area. The valleys are the mid-week days. Fort Lauderdale is definitely a weekend market.
We also have a strong high/low season differentiation. Everyone wants to come to Florida in the winter when it is cold up north. Not so much in the summer when it’s 95 degrees (35 C) and humid.
Of the three services, BeyondPricing is the most extreme. It sets the highest weekend prices and the lowest mid-week prices. Holiday and special event periods are similarly aggressively priced. In the low season prices are rock-bottom.
In the graph above, starting from the base price of $100, the highest peak is $209/night, and the lowest valley is just $55. There is a logic to this, and I will discuss it in more detail in a separate blog post (coming soon!).
Wheelhouse for a $100 Base Price
Wheelhouse is substantially less aggressive. The highest recommended price is $184, and the lowest is $84.
Note: the dotted line represents bookings that are already made and the price they were made at. I just put this apartment on the market a month ago, and I purposefully underpriced it to quickly build up reviews. Read about my pricing strategy for new units on Airbnb.
PriceLabs for a $100 Base Price
PriceLabs is a little more agressive than Wheelhouse, but not by much. The peak recommended price is $219, and the bottom is $83.
(At this point, another clear differentiator becomes apparent: Wheelhouse and BeyondPricing are substantially prettier. More on that later.)
Looking More Closely
So now that we’ve seen the macro picture, how the prices take into account seasonality and vary across the year, let’s look at what happens at the day-to-day level. Weekdays vs weekends is a big question.
If you are in an urban market with a lot of guests who come for short extended weekend vacations, then you probably find that your weekends fill up quickly with two- and three-night stays. But Sunday night to Thursday night often remains unbooked, or you have to discount heavily to get them booked. If this is happening, it’s an indication that your ratio of weekend to weekday pricing is not right.
All three services provide calendar views that let you see day-by-day pricing along with some explanation of the pricing.
BeyondPricing High Season Pricing
I’ve kept the base price fixed at $100/night. My apartment is in Fort Lauderdale, Florida. It’s a warm-weather winter destination for people who want to escape the cold weather in the northern US and Canada. The high season is approximately December 20th to April 10th.
Here are BeyondPricing’s price recommendations for next January:
Monday, Tuesday, and Wednesday nights are cheapest. Thursday and Sunday nights are about equal, and Friday and Saturday nights are the most expensive.
And yet there are oddities. Why is Sunday, Jan 12th so expensive? It is $165, when the previous and following Sundays are $133 and $136. Hover the mouse over the 12th, and we see there is a $30 increase for a local event. What event? That’s a mystery.
But if we look at Monday, the influence of the event has disappeared.
Wheelhouse High Season Pricing
So now let’s look at Wheelhouse for the same period. Unfortunately I couldn’t get more than three weeks on the screen at the same time, so we can’t see all of January. But the main themes come through:
- Much less agressive weekend vs weekday pricing
- A general increase of about 20% over the base price for all of January.
- Wheelhouse is unaware of this event on Jan 12th, but it thinks Jan 7th will be high demand.
Let’s explore what Wheelhouse has for Jan 7th.
PriceLabs High Season Pricing
The PriceLabs calendar is initially harder to interpret, but ultimately much more powerful. More on this later. For now, the sixes in the upper left corner are the minimum stay. For far-future stays I have a minimum of six nights. The upper left corner is the date, and of course the large number is the price.
While it lacks the wide range of weekend vs weekday prices of BeyondPricing, PriceLabs is across-the-board more aggressive. Prices are just higher. Let’s see what explanation PriceLabs gives us:
BeyondPricing in the Low Season
The low season is hard. I have a mortgage to pay, and in the low season I just hope to break even. A pricing mistake can be fatal because you can’t fill the calendar at the last minute by dropping the price a bit. To get the low season months filled, you need to be getting bookings steadily for months in advance.
Let’s see what recommendations each of the tools makes for September, which is traditionally the slowest month in Florida.
Tuesdays are priced at $55 compared to Saturdays, which are 63% higher at $80/81. Overall, prices are substantially lower than the base price of $100/night.
Wheelhouse is just flat across the board, with every night at $86.
PriceLabs has slight weekday/weekend differentiation, but not extreme. Other than October 6th, at $141, the differences seem to be in the 12-15% range.
Note, however, that September on average, is priced slightly above the base price of $100/night. This is strange given that September is traditionally the slowest month in Florida. Why is this?
I created the above screenshots at the end of January, approximately eight months before September. Anurag Verma of PriceLabs wrote to me to say,
Our algorithms apply something we call a “far out premium” to gradually raise the prices for far out dates (we’ve found this to be helpful in most urban locations).
As a result, the prices for summer/fall might be looking a little aggressive now, but will come down as the dates get closer. Same thing with Jan 2020 – the prices will come down as the month comes within the 2-4 month window (though seasonality will still keep them much above the base price).
This illustrates another key point about all of these products: how they work, and why they recommend certain prices, is pretty much a mystery. None of them publish long, detailed, insightful, analytic discussions of their algorithms. In fact, none of the publish much explanation at all.