PriceLabs’ Minimum Stay Recommendation Engine Algorithm

In 2017, we introduced our dynamic minimum stay settings, which quickly gained popularity. The ability to adjust minimum stay requirements based on lead time and automatically open up availability for shorter stays was a novel revenue optimization feature. It was a natural progression for revenue management systems, and we took the lead in introducing it.

In February 2022, we expanded these settings, offering even more flexibility with additional layers, adjacent night settings, and the option to set them differently for various seasons.

However, a common concern voiced by our customers was the challenge of determining the optimal settings. Questions like, “What should the minimum stay be for bookings far in advance?” and “How should these settings change based on lead time?” often perplexed users. Suppose you want to prioritize mid-term rentals—how should you adjust your settings accordingly?

Our data science team has been diligently addressing these questions since last year, and we’re thrilled to share PriceLabs’ Minimum Stay Recommendation Engine – World’s first and only intelligent minstay engine.

Why use dynamic minimum stay restrictions

Minimum stay restrictions are one of the key controls you have to create rules to decide which reservations you’re okay taking and which ones you’re not fine with. There are two primary reasons to use dynamic Minimum stay restrictions:

  1. Operational: The easiest example is that one-night weekend stays usually signal a party, which most short-term rentals want to avoid.
  2. Revenue maximization: Smart minimum stay automation can lead to a 5-10% increase in revenue
    1. If you set your minimum stay too low, you might get many short bookings far out. Taking a short booking precludes you from being able to take longer bookings on the days around it. So even though you got “guaranteed revenue” from the days that got booked, the dates around it now have a smaller chance of booking.
    2. If you set your minimum stay too high, you might be able to protect your calendar from getting long bookings, but any unsold days will see lesser demand, or if gaps are shorter than the minimum stay setting, none at all.

Our dynamic minimum stay settings allow you to change your minimum stay settings by lead time, gap length, adjacent days, and more! With this new update, we can recommend what those settings might be to maximize revenue!

The methodology: “opportunity cost” vs “guaranteed revenue”

At the core of our minimum stay recommendation engine is “opportunity cost.” In simple terms, selling a couple of nights 11 months out brings some “guaranteed revenue” (the revenue from those two nights). This feels great, and barring a cancellation, you are now guaranteed certain income for that month. However, for the dates surrounding the two nights booked, the chances of getting booked reduced pretty drastically. That drop in potential revenue from nights around the booked dates is the “opportunity cost.”

To illustrate this, consider the example below showing a calendar with 10 days and 2 nights (15th and 16th of the month) booked with a 2-night stay.

Let’s focus on the previous night (the 14th) and, for example, overlay the possible 4-night reservations that could book the 14th night.

Because the 15th isn’t available, the last 3 of those potential bookings aren’t really possible anymore.

Once the 2 nights (15th & 16th) are booked, it’s not just the 14th that experiences a drop in potential demand, but also other nights around it. For example, many week-long stays that would have previously been booked on the 11th will now be unable to.

The question remains – how many of these longer bookings could potentially bring larger revenue (by also booking the 14th and other adjacent nights) do we forecast in the market?

  • If not many longer bookings are expected, the opportunity cost is low, and we should take the guaranteed revenue from that short booking.
  • If any longer bookings are expected, the opportunity cost might be high enough to where there’s a benefit to holding out with a longer minimum stay requirement and not taking that guaranteed revenue!

The example above illustrates that one part of calculating the “guaranteed revenue” vs “opportunity cost” tradeoff is easy:

  1. Everyone knows the guaranteed revenue if you were to get that short booking.
  2. Estimating opportunity cost is trickier. Our data science team has been working on this since last year, and we’ve come up with a nifty way to incorporate local demand patterns and booking probabilities into an optimization framework to find that tipping point where taking a shorter booking will be worse for the overall listing revenue. That tipping point becomes your minimum stay recommendation.

With the above examples and context, you’ll notice a few things about our minimum stay recommendations:

  1. They tend to suggest having longer minimum stay requirements far out. Generally speaking, a lot of demand is yet to book, and you’ll still have a good chance of booking those nights later, even if you turn down demand for shorter stays.
  2. As a date gets closer (e.g., last minute), we’ll generally recommend reducing the minimum stay restrictions to take those shorter bookings. When there’s relatively little demand left to book, it’s better to take that “assured revenue” from that short booking instead of holding out for “potential revenue” from longer bookings that might not materialize.
  3. In markets where many longer bookings do happen last minute, you will see that the last-minute recommendations might not drop as much. This is because the recommendations look at the localized booking patterns of similar properties in the area (more on that below)!
  4. In markets with low overall demand, the recommendations would tend towards lower minimum stay restrictions (take whatever bookings you can get).
  5. High-demand months might be recommended in the seasonal suggestions for increasing minimum stay.

The demand data feeding our methodology

If you’ve been using our Market Dashboards in the last couple of years, you’ve probably already seen the kind of detailed booking pattern analysis that we have built over the years for markets around the world (the image below shows the demand by LOS for various stay dates in a ski market in the US).

When running the minimum stay recommendation algorithm for any listing, we look at demand patterns by length-of-stay (LOS) and booking window (BW) for various seasons for similar properties around your listing to suggest the minimum stay settings for your listings!

Length of stay patterns in an example Ski market (Big Sky, MT, USA)Length of stay patterns in an example Ski market (Big Sky, MT, USA)

Two modes: short-term vs mid-term rentals

Many of our customers (especially in urban locations) see a significantly higher proportion of mid-term bookings on their properties. The image below shows data for 2-bedroom properties in Chicago (our HQ!) – you’ll see that compared to the ski market above, Chicago sees a lot darker gray (15+ night stays).

Length of stay patterns in Chicago, IL, USA (an example urban market) show weekend heavy short term demand, but also a large portion of mid-term stays.Length of stay patterns in Chicago, IL, USA (an example urban market) show weekend heavy short-term demand, but also a large portion of mid-term stays

We created these two modes based on observations that many customers prefer one over the other for operational reasons.

  • If you select “Prefer short-term,” we’ll remove the 15+ night bookings from the market demand (and related supply) and show recommendations.
  • If you select “Prefer mid-term,” we’ll consider all the demand in the market. Note that even with all the demand included, the recommendations might still allow short stays if there’s not enough mid-term demand. Mid-term suggestions incorporate the demand for stays longer than 14 nights, but the recommendations might still be lower if such demand isn’t enough.

Stay restrictions in seasonal markets

For very seasonal markets (like the ski market above, or many beach markets), annual minimum stay settings do not work. Using our Minimum Stay Profiles in combination with Custom seasonal profiles.

However, the challenge for finding optimal and revenue maximizing minimum stay restriction for each season becomes even more complicated.

To help, we also run the opportunity cost optimization for each months’ demand in isolation to see if for a given month, the recommendations deviate from the overall recommendations. These “exception” months are called out with our recommendations, and you can create special requirements for these using Minimum Stay Profiles.

We believe that this will help reduce the amount of time you spend setting up stay restrictions and maximize income. If you have any questions about the Recommendation Engine, please do reach out to our support team and they will loop us in!

There are several operational reasons that you may want to consider while thinking about using our Recommendation Engine – we have made these recommendations editable so that you can use it exactly as you’d like!

Put your listings on autopilot by using our minimum stay recommendations and review them just once every month. As of end of 2022, Minimum Stay Recommendation Engine had been accessed over 40,000 times, so we believe we are on the right track!

Back to building,

PriceLabs Data Science Team

Get started with PriceLabs

Want to learn what PriceLabs can do for you? See for yourself with a free trial.

Discover our 2023 Milestone Release: The Breakthrough every Property Manager & Host Deserves


Join our newsletter

  • enter your email to recieve regular updates