How our Predictive & Reactive models are blended each day, to accurately price against both market prices, market bookings and your pacing.
In illustrating our Predictive and Reactive Demand Models, we were attempting to show that our pricing engine creates and utilizes two estimates of future demand, daily. The reason we have these two models is that we need to both (a) infer projected demand before it emerges, and (b) respond to actual demand, as stay dates approach and booking patterns emerge.
We leverage these two models for each day’s recommendation by dynamically blending the demand models together.
On the chart below, we show stay dates on the x-axis, from today (far left) to 365 days out (far right). On the y-axis, we show the weighting (from 0% to 100%) that our Reactive model has for any future stay date. The reason the curve slopes up (from right to left) is that as a stay date approaches, our Reactive Model starts to have a much bigger impact on that day’s price recommendation.
Additionally, the spikes you see in the line show where a large number of bookings cause our Reactive Model to be more heavily weighted sooner than normal. Said differently, when we see a big demand spike and have high certainty that this is a clear market signal, our model very quickly responds to that demand.
Dynamic Pacing Model
How groups of units can be leveraged to perfect pricing over time
While the last-minute discounting strategy detailed above works well for individual STR units, we have learned that for multi-unit properties (hotels, or premium STR buildings), an effective pricing strategy requires a much different approach.
To handle increasingly professional portfolios in the STR space, the Wheelhouse Pricing Engine developed the concept of ‘unit types’. Unit types help properties with similar rooms (such as multiple studios, 1BR, 2BR, 2BR/2BA, etc.) to leverage a ‘portfolio’ approach to pricing. Similar to how hotel pricing works, this approach allows us to make pricing decisions based on the performance of a set of units.
In the graph below, we illustrate the risk-reward ratio changes as a set of rooms begins to book. In this scenario, let us analyze a boutique hotel that has 10 units in a particular unit type.
Before the first booking occurs, the risk of a low daily revenue figure is high. Consequently, increasing prices prior to a first booking also increases the risk of the ‘worst case scenario’ — $0 in daily revenue. Alternatively, if 9 of the 10 rooms for a stay date are already booked, the marginal risk of increasing the price for the last unit is much smaller. Therefore, as more of a building or inventory type books, the more aggressively we can potentially price the remaining available units.
When we tested our pacing model on our portfolio through A/B tests, we observed an ability to drive revenue at a building level up by 0.9% to 1.5%. While seemingly small, impacts like this, in aggregate, are the foundation of our model’s success.
While one of our newer innovations, we are confident that Wheelhouse has developed the only dynamic pacing model in the STR category. We look forward to continuing to refine this, especially with our urban STR partners.