Property Recommendation Engine

[Elliott]

When I entered the first grade, I noticed lots of other kids knew how to tie their shoes. I asked my mother to teach me and I practiced over and over again until I mastered it. For the last 20 years there have been systems that have come forward and given me a list of comparables and I've had to write little comments as to why they selected horrible comparables. I'd say at a minimum, it would take a 'regular' person 2 years to figure out a market and get reasonably good at comp selection. The other complicating factor is that as field appraisers we get stuck with a small data set and have to make selections that are not necessarily 'obvious' to anyone other than an appraiser. I'll end with, when appraisers look at other appraisers reports, they frequently say, "I don't know why he used that one or those as comparables, the ones I used are much better." Its an art, not a science.

 

[RealtySort]

Hi Elliot. I agree with you, selecting comparables is definitely an art. Yes, there is lots of subjectivity based on the market and the assignment. As an appraiser like you, I'd like more objectivity and quantifiable adjustments in my reports too.

How about we put the system up to a challenge? Let's put some comparables it to the test and see how well it does. Let me know if you're interested. I can send you a list of comps and a subject property and you rank the top 10 or feel free to send me a list of comps and a subject property and we can put it through the system to rank the top 10.

 

[RealtySort]

Hi guys,

I'm excited to tell you that we just launched Realty Sort, our comparable recommendation tool for appraisers.

I’ve put together a 1 min video to show you how you can use it:

Let me know if you're interested in using it for free in your next assignment.

Shawn

 

[CANative]

Arggg...Noooooo!!

This may be well suited for the cretins that wear penny loafers and do 20 appraisal a week. For a real appraiser this is sort of offensive.

 

[Terrel L. Shields]

. I'd say at a minimum, it would take a 'regular' person 2 years to figure out a market and get reasonably good at comp selection.

I'd say that "regular" person was a slow learner. I am not saying that we don't get better (and/or faster) with time, but I am saying this ain't rocket surgery....

Our local vendor of assessor records has a search feature that does much the same thing...

 

[Meandering]

The MLS does exactly the same thing.

The problem is, and has always been, the crap in the MLS gets picked up by nobo programmers as if it's REAL data, and then appraisers don't find the "right" comps because of a screwed up check box or zip code, in the MLS data.

Well at least he joined the forum to give us something free.

Hey RealtySort, you want to program something appraisers need and will buy, email me and we'll work a licensing agreement out.

.

 

[George Hatch]

Here's the thing; it looks like your tool is built to sort the data for the appraiser with the goal being to narrow down the bigger dataset to the smaller dataset used for direct comparison.

The underlying premise seems to be that the appraisal hinges primarily on the analysis of the direct comparables and that any process that gets the appraiser to that smaller dataset more quickly is a good thing.

My problem with the shortcut is that my process goes from analyzing the subject within the context of it's immediate environment >> analyzing it within the context of its market segment >> and finally, analyzing it in direct comparison to the sales on the grid. I don't want to analyze those direct comparables in isolation to all the other data. It's not my direct comparables that comprise the market, it's all the data.

So my appraisals don't hinge on an analysis of only the 6 sales that appear in the report. Before I narrowed it down to the 6 I put into my report I have ALSO analyzed many other sales and am using those to provide the background and context for the sales I am presenting as my direct comparables. I commonly look at actives, expireds and withdrawns to see what listings failed in the market and as an indication of what the subject most likely isn't worth.

The sales I present in my report are the tip of the iceberg of data that I analyzed - to one degree or another. They're indicative of the underlying trend; they do not comprise a separate trend of their own. So I don't want to skip the step of the macro analyses of the raw data. I'm not looking to be an AVM-Lite and I'm not trying to be the calculator for working a 6-record dataset. I WANT to start from the outside and work my way in via the logical progression. I WANT to be more familiar with the macro data than any of the other participants in the transaction.

Just sayin'

Maybe your target market is the form monkeys who just want to put out completed report forms with their signatures on them in mass quantities.

 

[RealtySort]

George,

I agree with you, it's still critical to vet the qualitative data with pictures and full MLS brochures to find the most relevant ones. We're providing a tool to help facilitate the process of narrowing down hundreds/thousands of potential properties (actives, sold, expired, etc..) to find the top ones for vetting. All of the data can be uploaded which eliminates the isolation challenge you mentioned.

 

[Dale Smalley]

After the Location,Location, Location thingy I am down 30 -40 sales and listings Then I just resort the lists by time then attributes like GLA, age lot size etc until I have about 10-15 properties to look at in the field which the best 6 are used. Believe me I am always looking for good comps with good data.