You might not be familiar with the term “recommendation engine” but you’re probably very familiar with the outputs of multiple recommendation engines because they’re quite common in our everyday life.
Tonight, when you’re ready to relax after sending off that last email and putting the kids to bed, you won’t have to think about what to watch. That’s because Netflix’s recommendation engine has already picked a few movies and shows for you.
And when you log onto Amazon, you’ll be tempted with “similar product” recommendations based on the items you’ve previously purchased or the ones currently in your cart, all thanks to Amazon’s recommendation engine.
Recommendation engines try to predict what you might like based on your previous choices and compared to other people with similar interests. From movie selections and product suggestions to online advertising and food delivery, recommendation engines surround us in all manner of situations.
Here at Archer, we’ve brought the magic of recommendation engines to commercial real estate. We’ve built a proprietary recommendation engine that identifies new areas for our investors to investigate and properties for our clients to pursue.
Creating a “digital twin”
Recommendation engines have been used extensively in the world of single-family homes and apartment rental searches. Anytime you complete a search for a house for sale and you receive a “similar results nearby”, this is a recommendation engine at work. However, in commercial real estate, and specifically in acquisitions and investment analysis, recommendation engines are quite rare.
Archer’s recommendation engine is unique to the firm because of the extensive work we’ve done to identify, capture, and normalize the data needed for this type of model. Our team of talented data scientists originally built our recommendation engine to help our clients expand from their local market into adjacent markets where they have a less of a presence.
Archer is currently evaluating more than 100 markers, looking at those markers over time and at a static point in time to generate recommendations for our clients. In addition, we go a step further to analyze and evaluate properties that are likely actionable and align with our client’s investment strategy. The methods we use to capture our client’s investment strategy to effectively create a “digital twin” of their investment strategy is unique (as far as we know) to the industry.
Making market entry easier
The primary application of Archer’s current recommendation engine is to helping investors enter new markets easily. By understanding investor preferences in a given geography and the investment strategy, we can build robust recommendations in any other market for our clients to pursue. Our powerful recommendation engine can analyze new markets at scale that would traditionally require an army of analysts.
Recently, Archer recently partnered with a Houston-based private equity firm to expand its investment searches to Austin and San Antonio. We are also working with a client who has a very specific use case in San Diego and Salt Lake City, and we are looking for specific pockets in those markets that align with their investment criteria.
All recommendations generated by Archer’s recommendation engine are vetted by our internal experts. We use a monitored recommendation system and work closely with our clients to validate the recommendations through numerous feedback loops.
Each iteration of the model that we run for a client improves and refines the output, improving the quality of recommendations with each iteration. In some cases, we can even identify where our client’s preferred locations may not align well with their stated strategy. In these circumstances, Archer helps our client make subtle shifts to their location preferences based on the data at our disposal.