All Forum Posts by: Johann Villalvir

Why would one need an app? If you have experience in the area you should know, well you better know the ARV. Any property any street in all of my markets I will tell you the ARV within a few bux in under 5 minutes. Of course, some will be a bit higher or lower, depending on upgrades, appliances, tile, crowns, flooring etc. Also why not just run comps? IMO this does not make any sense, but what do know. So many are trying to reinvent the wheel when its just about numbers.

All the best 

Hello @Johann Villalvir,

Congratulations your accomplishment.

Some background: I am an engineer and what I do in our investment services business is primarily data science and software development.

We use data mining software to evaluate thousands of properties based on about 40 behavioral characteristics of our target tenant segment. The software performs the following functions:

• Eliminates all properties that do match our target tenant pool segment.
• Selects sales and rental comps for evaluating rent and return.
• Generates investor analytics. We provide actionable information, not MLS data sheets.

I’ve experimented with increasing the scope of our software. However, despite trying many approaches, I have failed to consistently meet our accuracy and consistency requirements.

Today, we use a hybrid solution. We reduce the number of investment properties to consider by about 99.6% using our data mining software. Then, we match the properties that passed the data mining engine against the specific financial situation and goals of individual clients. Next, we manually evaluate each property. At the manual evaluation step, approximately 60% more are eliminated. Additional software estimates renovation costs and provides actionable analytics. We have never shipped MLS data sheets.

Below are some reasons why taking the next step with software, has not been successful for me.

• There is a lot of inaccurate information in the MLS. A recent example, "How much is the monthly association fee?" Answer in the MLS, "Yes."
• Computers cannot accurately determine whether a property is located in a desirable or undesirable area for a particular tenant pool segment using photos or aerial maps. There are more location considerations. For instance, even if a property has excellent characteristics, it may not rent if there is a significant amount of traffic on the street. This determination can only be made if you know the specific street.
• An onsite evaluation is essential. For example, if there is a barking dog next to a property, it will not rent. How can software determine the presence of a barking dog?
• Zillow and OpenDoor attempted to use software to predict the sale price of properties and their results were/are bad because their software was poor and the inherent limitation of software to deal with subjective data. Zillow lost so much money that they decided to exit the flipping business, while most of the OpenDoor properties we see in Las Vegas are sold at a loss.

In my experience, software can only go so far when humans are involved. Our hybrid approach has worked well. If there is a better approach, I would like to understand what it is.

Thanks for your post.

Hi! for what you’re asking, a free internet based platform called Homestyler is enough to study concepts like floorplan layouts, furniture space planning, and creating nice looking renderings.the learning curve is pretty easy. SketchUp has a free version and it’s come a long way. The Pro paid version can create a decent set of plans, with elevations, details, finish schedules etc but for the average investor, Homestyler is my recommendation. there are many out there though.  Below is a before and after of a studio condo I’m considering of purchasing to attract traveling nurses that I used Homestyler to quickly model out the space.