Suggests for "Comp" software

[Mel Park]

Happy New Year all....

Curious if anyone can suggest the right software for figuring out comps.  Yes, I can guesstimate based on previous sales, etc --  but sometimes the same house in the same zip code - can sell for vastly different sums depending on what part of town it is.  Is there software that is mostly accurate for comps?

I'd like to buy a rehab home -- but want to get a pretty accurate read what the house is worth if made ready for sale.  Thanks

[Nathan A.]

I’m sure software is getting better every day but this is too important a number to trust to an algorithm. I would do it yourself so you can “show your work” and trust the conclusions. There simply is too much data that goes into valuing a property that isn’t clean and “machine readable” as they say in the software industry, unless you’re in an area with a huge supply of genuine cookie cutter houses.

If you’re not sure why some houses sell for a lot more than others I would find a local agent for your team who can give you insight.

All of that said, there’s Zillow Zestimastes if you just want a general idea.

[Frank Greg]

I usually advice against this. Its best to be in a position to operate with or without an app or software when trying to estimate the value of a potential property or deal. You can obviously use the app or the various websites to confirm or check your work.

To expect an app (as useful as some may be) to anticipate every nuance or unique characteristic associated with every property, terms of sale, condition of property etc. to generate an exact valuation can be a reach.

There usually can be drastically wide swings between neighborhoods often with the same postal code that the median or average for a city might be waaaaay off when looking at certain pockets within a city. The app is only as accurate as what goes into the algorithm and the data source.

[Chris Seveney]

@Mel Park

Propstream we use for comps as it gives us sales data for the area where we can determine comps

Datatree has a ton of AVM’s you can run as well

Both cost $

[Eric Fernwood]

Hello @Mel Park,

The problem is not software. You do not need more than a spreadsheet. The problem is selecting the right comps. The inability to select the right comps is why all the major real estate websites fail to produce reliable numbers.

First, I will show the process. Second, I will talk about selecting good comps.

Suppose you are considering the following property.

You find two recent comparable sales, which are listed below.

The process is to determine the average $/SF of the comps and multiply that times the subject property’s SF. (238/SF + 235/SF) / 2 = 236.5/SF. 236.5/SF x 1847SF = $436,815. The condition of the property can have a significant impact on the market value.

Rental comps are handled the same way. Below are three recently rented comparable properties.

Calculation: (1.05 + 1.03 + 1.11) / 3 = 1.06/SF. 1.06/SF x 1847 = $1,957

The hard part is selecting comparable properties. This is where experience is required; no software can select good comps. Software does OK with objective data (SF, bedrooms, lot size, etc.) but fails with subjective data like photos (condition), proximity to nuisances, etc.

Below is a summary of the comp selection guidelines we follow.

• Distance: All comps should be within 1 mile of the subject if it is in an urban or suburban neighborhood.
• Age: All comps should be within ten years (newer or older) of the subject property.
• Size/square Footage: Only above-ground square footage is calculated in the square footage. The square footage on the comps must be within 20% of the subject. Basement square footage and finish are calculated separately from the above-ground square.
• Property type: Comparable properties must be the same as the subject property. Single-family detached homes must be compared to single-family detached, duplexes to duplexes, townhouses to townhouses, etc.
• Bedroom/bathroom count: Only above-ground bedrooms and bathrooms are counted. Comparable bedrooms and baths count must be within one of the subject properties.
• Style: It is not required, but you should try to use the same style of home. 2 story to 2 story, split level to split level, ranch to ranch.
• Sale date: Sold comps should have sold in the last three months.

The above are guidelines because you will likely not find ideal conforming comparables. An (experienced) human can make adjustments based on the best available comps. But no software can make that kind of adjustment.

There is another consideration with rentals. While recent similar rentals are good indicators, the current competition is more important in determining the rent. For example, suppose there are four similar properties in similar condition for rent nearby. If you want to rent your property quickly, you will likely price your property lower than the competition. The competition determines how much your property will rent for.

[Eliott Elias]

It's not guesstimating, the only way you can run comps is pulling recent sales data. 

[Frank Greg]

 There are of course pros and cons with this. The more you drift toward the realm of the "subjective", that can also be a problem as the result becomes more vulnerable to what often can be substantial manipulation if the realtor for instance has an interest in the transaction.

If software has any advantage, there is some value in knowing what the average or mean is for an area based on several hundred or thousand+ transactions as opposed to 10 or 12 properties (or less) in a manual computation. 

Most apps or website can also enable the user to quickly get a ballpark feel/pulse for an area -- which can be helpful if you have thousands of potential deals to sift through. 

[Eric Fernwood]

Quote from @Frank Greg:

 There are of course pros and cons with this. The more you drift toward the realm of the "subjective", that can also be a problem as the result becomes more vulnerable to what often can be substantial manipulation if the realtor for instance has an interest in the transaction.

If software has any advantage, there is some value in knowing what the average or mean is for an area based on several hundred or thousand+ transactions as opposed to 10 or 12 properties (or less) in a manual computation. 

Most apps or website can also enable the user to quickly get a ballpark feel/pulse for an area -- which can be helpful if you have thousands of potential deals to sift through. 

 Hello Frank,

You are correct; software can give you an average price for an area, as you see on Zillow and other sites. However, such software cannot provide accurate (or even reasonable) information on a specific property. Below are some examples.

Software is good with analytical data (SF, beds, baths, etc.) but fails with subjective data. For example, suppose you have two properties. One property is in exceptional condition, and another is in poor condition. Software will determine the average $/SF for the area and multiply this by the subject property SF. This means that the predicted price will be low for the property in exceptional condition and high for the property in poor condition. Software fails in other ways too.

Are the comps below good comparables to the subject property? The answer is no. The properties are similar in size and such, but the subject property is next to a freeway, which is very loud. The comps do not have the freeway noise. In such a case, the subject property might look like a bargain, but in reality, it might be overpriced.

Below is another example. The subject property is in a community where the prices are about $180/SF. Across the road is a completely different community, with prices ranging between $280/SF to $310/SF. Because the properties are in close proximity and similar in size, they will be included in the price calculation for the subject property, which is invalid. So the subject property might appear to be a bargain, even if it is overpriced.

Online software can tell you the average $/SF for an area, but provides no useful information for any specific property. Never evaluate a property based on Zillow, Rentometer, Redfin, or any other similar source. The only way to evaluate a property is to have a knowledgeable person evaluate the specific property based on similar recent comps, taking into account location and condition differences.

[John O'Leary]

I use propstream and batch leads. 

[Mel Park]

Quote from @Eric Fernwood:

Hello @Mel Park,

The problem is not software. You do not need more than a spreadsheet. The problem is selecting the right comps. The inability to select the right comps is why all the major real estate websites fail to produce reliable numbers.

First, I will show the process. Second, I will talk about selecting good comps.

Suppose you are considering the following property.

You find two recent comparable sales, which are listed below.

The process is to determine the average $/SF of the comps and multiply that times the subject property’s SF. (238/SF + 235/SF) / 2 = 236.5/SF. 236.5/SF x 1847SF = $436,815. The condition of the property can have a significant impact on the market value.

Rental comps are handled the same way. Below are three recently rented comparable properties.

Calculation: (1.05 + 1.03 + 1.11) / 3 = 1.06/SF. 1.06/SF x 1847 = $1,957

The hard part is selecting comparable properties. This is where experience is required; no software can select good comps. Software does OK with objective data (SF, bedrooms, lot size, etc.) but fails with subjective data like photos (condition), proximity to nuisances, etc.

Below is a summary of the comp selection guidelines we follow.

• Distance: All comps should be within 1 mile of the subject if it is in an urban or suburban neighborhood.
• Age: All comps should be within ten years (newer or older) of the subject property.
• Size/square Footage: Only above-ground square footage is calculated in the square footage. The square footage on the comps must be within 20% of the subject. Basement square footage and finish are calculated separately from the above-ground square.
• Property type: Comparable properties must be the same as the subject property. Single-family detached homes must be compared to single-family detached, duplexes to duplexes, townhouses to townhouses, etc.
• Bedroom/bathroom count: Only above-ground bedrooms and bathrooms are counted. Comparable bedrooms and baths count must be within one of the subject properties.
• Style: It is not required, but you should try to use the same style of home. 2 story to 2 story, split level to split level, ranch to ranch.
• Sale date: Sold comps should have sold in the last three months.

The above are guidelines because you will likely not find ideal conforming comparables. An (experienced) human can make adjustments based on the best available comps. But no software can make that kind of adjustment.

There is another consideration with rentals. While recent similar rentals are good indicators, the current competition is more important in determining the rent. For example, suppose there are four similar properties in similar condition for rent nearby. If you want to rent your property quickly, you will likely price your property lower than the competition. The competition determines how much your property will rent for.

[Mel Park]

@Eric Fernwood

Wow, I really apologize. I distinctly remember writing a rather verbose reply way back when - - and today I see there's no reply from me so don;'t know if I didn't hit "post reply" or what might have happened.

Regardless, thank you very very much for that insight and detail, very appreciated.  

[Eric Fernwood]

Hello @Mel Park,

You are welcome. Rereading my post, I want to add some additional information and it has to do with rental properties.

Rental comps are different than sales comps. People think that their rental property’s competition is the property for rent down the street. That is not true. Renters select rental properties based on proximity to where they work. See the image below. The competition might be across town. The ability to recognize the competition for a property is a skill few property managers possess.

In the current Las Vegas market, there is a strong demand for rental properties, as evidenced by falling inventories and more homes rented (see below). However, it can be difficult to determine the correct rental price. One strategy we use is to ask for a higher rent than the estimated rent and then monitor the showing traffic. If we don't see the expected number of people viewing the property, we reduce the rent. This way, we maximize the rent and capture potential tenants' attention in the first few weeks on the market. Also, reducing the rent triggers notifications to people who are on drip feeds for a rental property.

The following charts are for our target property profile only.

The number of homes for rent is decreasing.

At the same time, the number of homes rented is increasing, indicating increasing demand.

The result is less than 1 month of supply for rentals. This will push up the rent.