Dynamic Pricing Innovation: The Evolution of Atlas at Bukit Vista

img Furqon Bukitvista | August 21, 2023

Greetings! I am Furqon, a dedicated and results-driven PR specialist with a passion for crafting compelling narratives and building strong relationships. With a keen understanding of the technological implementation in the hospitality industry, I am excited to share our latest update about our property pricing recommender.

In our relentless pursuit of innovation, we’ve embarked on a groundbreaking journey that fuses cutting-edge technology with hospitality expertise. Introducing Atlas previously known as Promenade, a remarkable tool that transcends conventional revenue management. At Bukit Vista, we’re not just redefining the hospitality landscape – we’re revolutionizing it, one data-driven decision at a time.

Atlas: Where Innovation Meets Intelligence

Atlas is a system that collects data in the form of available blocks from rental properties in Bukit Vista, it then suggests a price change and provides the revenue management team insights to approve or reject the suggestion and update the price to OTAs (e.g. Airbnb)

Atlas  goes beyond mere suggestion – it’s an orchestration of data-driven intelligence. By calculating a multitude of factors, from seasonality to booking count, runway to current rates, it crafts the Suggested Price Event Decision (SPED).
SPED is a proposed new price setting and is reflected by percentage.

 
image.png

Example: The current price (previous rate avg) for Asri Village is at IDR 610,000, with +3.7 SPED, Atlas suggested that the price should be increased by 3.7% → IDR 632,570.

But here’s the twist – this suggestion often dances to its own rhythm, distinct from the Pricing Event Decision crafted by our adept Business Intelligence team.
image.png

Continuing from the previous example: Our Business Intelligence team decided to lower the rate by 15% – reflected as -15 in the Price Event Decision (PED) row → resulting decrease to IDR 520,000.

Atlas Impact

As a result, Atlas helps our revenue management team to price each date of the calendar by getting and collecting data like booking window distribution, pricing distribution, runway, etc. Afterward, the revenue management team prices the dates in the form of a change in percentage which we call PED (Pricing Event Decisions). Which we keep on doing until we reach enough data to train a machine learning model to predict it for us.

However, this sparks a big question, how to detect which prediction was right?

Felix’s Trailblazing Vision: Atlas Model Explainer

Cue Felix, the luminary of our Data Engineer team. Sensing the harmonious potential between technology and human insight, Felix embarked on a visionary project: the Atlas Model Explainer visualization. This marvel of innovation peels back the layers of decision-making, showcasing the intricate factors that coalesce to shape the Pricing Event Decision.
Meet Felix, our agile Data Engineer who build the model

Pain points of previous version of promenade

Before, Promenade lacked vision as it was outputting random and wrong PEDs. We would like to know what causes this, as the Machine learning model learns from data, we would like to see how each instance is being predicted and why the model predicts a certain PED. The model becomes something like a black box without any explanation of why.

On August 15, a new epoch was ushered in – the launch of the Atlas Model Explainer. With this unveiling, our Business Intelligence team has newfound insights into the factors at play in the outputs of PEDs. The realization dawned: decisions could be based not on assumptions, but on the foundation of intricate data.

 
image.png
 

From the picture, we can see the contributing factors for set PED (-7.25%). One thing that was significantly noticed by the Business Intelligence team in this specific case was that they over-weighted the current price factor (represented as prev_rate_avg) too much. It dragged the potential revenue numbers that we could achieve.

The inner workings of the explainer

We use LIME (Local Interpretable Model-agnostic Explanations) which uses local prediction to explain the model. It generates perturbations of data inputs those data to the model and sees the output. As we increase and decrease the input data of each feature, we will be able to see what has a high and low impact on the predictions.

The Atlas Model Explainer helps us understand a complex model easily. It’s like opening a black box and getting a clear explanation. This allows for faster feedback to improve the model. Moreover, it strengthens data quality and makes analysis easier through improved logging when saved in a database. With this, we are able to detect some features like the previous rate average (average price for a given available block) doesn’t make sense because it doesn’t align with the revenue management analysis to take a certain decision. 

Charting New Horizons: The Call for Data Scientist Visionaries

As we stand at this juncture, the echoes of innovation reverberate across Bukit Vista. The Atlas journey has just begun, and we’re inviting the brightest minds in data science to join us. In the future, we can enhance the system by adding more visuals to the dashboard, making it easier to interpret data. Additionally, incorporating algorithm-based A/B testing methods is an option to improve testing and decision-making processes.

Excited? Be part of a revolution that’s rewriting the playbook of data-driven hospitality. Learn more about what we do here at Bukit Vista, discover the future of revenue management, and chart your own path with us.

Compare listings

Compare
💬 Need help?