Sellsius real estate marketing blog

thanks your commetns,this is good job!It was a very nice idea! Just wanna say thank you for the information you have shared. Just continue writing this kind of post. street lighting street lighting christian louboutin discount .I will be your loyal reader. Thanks again.

Thanks Lar.
Shortly after this debate Zillow did 2 things (1) allowed owners to add, correct and supplement data (an admission that zillow's data from the public record was not doubled checked for accuracy and completeness as David claimed was done by the tax assessor)
(2) displayed an owner's calculated estimate. This was in response to our complaint that the my zestimator tool results were not made public.

Yet despite the changes the zestimate still lives. Will the failure to kill the zestimate (in some way),lead to the death of zillow as a realtor listing site?

I wish that everyone in this USA could have the opportunity to read this point and counterpoint. David struggles, in my opinion, and loses the debate. Lar

http://www.washingtonpost.com/wp-dyn/content/ar...

More food for thought.

I think I will use this comment to bring things full circle to the unzillowable.

Where we diverge is that you say they're already implied in the SALES data----and that may very well be for those SOLD properties. (this is still problematic because you cannot know from the comps which unzillowables were attached to which sold house). But this is NOT what we are talking about when we discuss unzillowables. We are talking about MY unzillowables for my UNSOLD house. The comps, we argue, do not factor in MY unzillowables. Here's what we mean. There was a comment in your blog (which, BTW I commend zillow for) by a man who said he had a unique (one of a kind) view of the ocean and of all the houses on the beach, his was the only one that did not have mud when the tide went out. That's unique to his home. We would say these are HIS unzillowables & they are not accounted for in the sold comps because no other comp had those unzillowables. We hope this example helps explain the distinction.

Also, the fact that ohhs & ahhs (unzillowables?) are factored in post facto using the zestimator tool, ie, the tool is used AFTER PUBLICATION of my zestimate, means that the original published zestimate DID NOT FACTOR THESE OOHS & AAHS . (I am not shouting it's just I cant make bold print).
The tool is fine but we question the logic of keeping the results of such a tool private to the owner. We think that is a disdvantage to a buyer & leaves a possibly misleading zestimate on the public table.

Q5: Have you researched whether the laws in some states regarding conspicuous "non appraisal" disclosures apply to zillow? I just dont know but I assume someone has done the research. Is it really apples and oranges?


Re: starting points

Almost anything can be a starting point, even my neighbor's opinion. But we want to get the BEST starting point. Here we diverge again. We say there is a better starting point than zillow. In our opinion you're better off using an experienced professional who comes to your home with fresh, accurate & complete raw data (based on their review of that raw data) and, to use terms from the Inman article & Jonathan Miller, "uses human experience and complex thought processes to evaluate some factors the computer cannot access. The value of these features is often more intuitive rather than quantitative." " Professional are needed to bridge the gap between raw data and subtle nuances within a property or market that are not picked up in the data. "

In conclusion, we support consumer choice, including the right to choose zillow. Good luck and thank you.

Joseph

Hi Joseph;

I feel my time has been well spent discussing these details. Your questions were answered in my previous comments so, yes I felt repeating them was not necessary. Sorry I misunderstood your reference to Zillow's business model.

I'm glad you found my answers. Here are some further clarifications where necessary:

Q 2. - we will fix bugs in our data and Zestimate processes which is why user feedback is so vital. This is not the same as letting users decide that their Zestimate should change or what it should be. Is that distinction clear?

Q 3. No, they include Zindexes. Refer to my comment above where I explain how outliers in Zestimates are effectively ignored by this metric.

Q 4. You've defined a fictitous issue based on incorrect assumptions about how AVM's work. What you define as "unzillowables" ignores implied attributes (see above) and I'm not sure how to make this issue clearer. Do you understand that sales prices reflet the value of the attributes you consider "unzillowable" or do you really believe that buyers ignore them?

Q 5. The fact that a Zestimate is not an appraisal is clearly indicated on the site.

I urge you to spend more time reading the content on Zillow. A brief read of the site would clear up much of the confusion around Zestimates. Watch Zillow in the near future for a better idea of how our product will improve and evolve.

No irony here; I was commenting on your blog's ongoing series on so-called "unzillowables", mentioned yet again in this post. Given the time you've invested in this topic, I thought that it made sense to weigh in with explanations where necessary. Hopefully they've been useful.

David

I apologize if you feel it is a waste of your time. I don't feel it's a waste of mine. Some things bear repeating. Sometimes the answers are not clearly apparent.

Counter-Correction (is that a word?):
My statement was "We (sellsius) are NOT an advertising model like zillow...

Sorry you misinterpreted this. We know zillow is not subscription based. Zillow makes it's money from advertisers.

We also know you have a range & you say the wider the range the less reliable the zestimate. So in a very wide range, the zestimate's margin of error is also likely to be very wide. Yes, it is user facing, but I don't think a very wide range is a cause for confidence in that zestimate.

If you noticed, I numbered the questions in the hope you'd answer them that way. There were only 5. I didn't hear them answered directly but I think they are as follows:

1. Will Zillow let Owners Opt-out of publication of their zestimate for any reason, "even if grossly incorrect"?: No. Based on your statement: "We do not let people...contact us to remove zestimates"

2. Will zillow hear an owner's appeal (request)to correct their zestimate?:Yes. Will zillow then correct it? Yes
Zillow will hear an owner's objection & correct a zestimate. I cite your statement: "We found the problem with Mr Olshin's zestimate & hope to correct it soon".

3.Do zillow's house reports for a given market contain comps of sold houses together with zestimates of unsold houses?: Yes.

4. Do you admit to anything being "unzillowable" (as we define it) and not factored into MY published zestimate?: Yes, by virtue of the zestimator tool. Will the tool-adjusted zestimate be published? No

5. In those states that require it, should Zillow's zestimate be held to the same standrard as a Broker's CMA & conspicuoulsy post a "non-appraisal"disclaimer on the HP & zestimate?: No


We have been reading and listening and that's why we asked these questions. We're pleased you thought some were good. I'm sure you'll hear the questions again from others, along with better ones

Best of luck to you and zillow. Unfortunately, none of us can determine our greatness, it is for others to judge.

Best regards,
Joseph Ferrara

PS: I find it ironic that this discussion was triggered not by Sellsius' words, but by the words contained in an Inman article and a post highlighting same in a renowned appraiser's blog, Soapbox by Jonathan Miller.

Hi Joseph;

Dialogue is good and all, but it is not a good use of your or my time to repeat ourselves here or to debate inaccurate statements. Please, read my comments above, and visit our website -- you will notice that your questions are answered and that neither of these statements are true:

"Why do you feel a need to provide a precise number rather than a range or just the raw data & let others decide themselves what the number should be?"

Correction: Zillow does provide a Value Range for every single Zestimate - the size of that range is another user-facing indication of Zestimate confidence.

"... like zillow–we are member subscription based"

Correction: Zillow is not "member subscription based".

You've asked a few good and some subtle questions -- now I must respectfully ask you to "listen" to the answers. Keep an eye on Zillow.com -- like all great internet companies, it's a work in progress. As a visitor, you should notice a lot more progress in the coming months.

Best of luck with your venture. Please let me know when it launches.

David

Thank you for your comment David. I appreciate the dialogue.

Regarding the assumptions:

1.Market conditions/market pool affect pricing more than a zestimate

I think current market conditions, as well as the current market pool of homes for sale, affect the price for an unsold home, more than it's zestimate. An example: In NYC during the rising market, you could not buy a property at list price --you had to offer ABOVE list price. How much above?---your computer could not tell you. If you were a buyer in that market I don't see how the zestimate would have helped you, even as a starting point. Knowing the value, even if 100% accurate would not have gotten you a house in that rising market. A poor analogy perhaps is trying to buy Google stock 1 day AFTER going public at the IPO price. You'd be too late & not get the stock. Your public data is, for the most part, not up to the day/week/ or even months fresh. That matters in a rising market.
I suspect the zestimate would also lose value in a rapidly declining market, from the seller's point of view. You might be pricing too high as a result of using a zestimate based on the past rising market comps.

2. To be valuable data must be fresh, accurate & complete

This I think is beyond dispute. And these cannot be known without human checking beforehand. The fact that there are gaps in zillow data (beds, baths sq. ft missing) tells me that that data has not been checked.

Re Completeness: I have examined zillow's data & even though beds, baths or square footage may be missing, you will still publish a zestimate.
That affects accuracy. It would be better if you didn't publish a zestimate if there were incomplete data. Don't you agree?

Re: freshness: Public sales data is not obtainable the day of sale. In some states it may take months. In a changing market the zestimate will lag.
Also, as you know, tax assessments are not done every year. Thus, where you are using tax assessments, which may be several years old, you are using stale data.

Re: accuracy: The zestimator, as you point out, is a tool to "correct" (adjust) accuracy. So if I fill in (or correct) the beds, baths, square footage, the zestimate is rendered more accurate--fine. The only flaw is that once it is corrected with this tool, the new zestimate is NOT PUBLISHED by zillow. Why is it not published? That leaves a possibly greatly flawed zestimate of a person's home on the public table for all to see. You may argue: "well the owner sees it"---but the BUYER doesn't.

Human double checking & a visit to the tax assessor's office.

As an attorney, I have succesfully challenged several tax assessments of value for property I own. As you know, tax assessors DO NOT GO IN A HOUSE. In some cases they might rely on a drive by. More commonly they rely on sales data, comps. They are crunching numbers just as you are. But zillow is at a disadvantage. As I mentioned above, it is important to point out that tax assessments are NOT recalcualted EVERY YEAR. Thus zillow is likely using stale tax assessment ingredients in their sauce. (sorry for the pun--is that a pun?). In MD, property is reassessed every 3 years. In PA, some counties have not been reassesed for many more years.

Nonetheless, the tax assessor values are APPEALABLE by owners. Zillow's zestimates are apparently not appealable. Which brings me to the next points.

I am happy to hear Zillow is correcting Mr. Olshin's Zestimate. Am I to assume from this that Zillow will correct other grossly inaccurate zestimates, if another owner asks? Please answer this David.

Perhaps I missed it, but I do not think you answered these questions:

1. Since the zestimate is proprietary & not public data, will you allow an owner to "opt-out" of "publication" of his zestimate for any reason? Let's say they ask you to take it down because it's grossly inaccurate? Will zillow honor the owner's request in that case? Evidently not.

2. Can an owner appeal (ask you to correct) his zestimate so that you change it online where potential buyers can see it---like what Mr.Olshin is in essence asking you to do?

3. Do your house value reports include zestimates as well as sold comps? If they do, won't zestimates like Mr. Olshin's be in the calculations? Will it not affect the conclusions in the report?

Re: Medians values of zestimates to offset skewing caused by inaccuracies.

As anyone who studied statistics knows, the median is just the the middle value. So if zestimates are 200K 210K 400K 800K 850K the median is 400K (let's say that's Mr. Olshins). The average is $490 Is the median value better than average value? The difference between the 2 is over 20%. Or are both wrong becaus Mr. Olshins is not accurately reflected? Furthermore, how do you know that ANY of the other zestimates are likewise incorrect, & to what degree?. You can't KNOW until the homes which these zestimnates are atttached to are ACTUALY SOLD. Thus merely taking the median value of "zestimates" does not address the issue of the accuracy of MY zestimate. Accuracy is a post sale conclusion/calcuation---this is how zillow's error rates are determined, correct?

4. Do you admit to anything being "unzillowable" as we have defined it? If so, what is currently unzillowable?

5. Re: disclosure. The point here is that experienced brokers, knowledgeable and trained in real estate, are required in some states to attach a CONSPICUOUS disclosure to their CMA, while zillow "statisticians" are held to a lower standard. You do agree that brokers are held to a higher standard? We think zillow should voluntarily post a "conspicuous" disclaimer on the homepage & beside each zestimate before a state like MD, NJ or PA imposes the requirement. It is the higher road IMO.

I also need to ask zillow this? Why do you feel a need to provide a precise number rather than a range or just the raw data & let others decide themselves what the number should be?

I guess the bottom line for us is that THERE ARE in fact UNZILLOWABLES that are not factored into the published zestimate of your home & since you cannot know until AFTER the sale of your home how far off the zestimate is, the BETTER starting point is using an experienced professional to come examine your home and its environs to factor in the unzillowables WITH the published RAW DATA, which is the sweeter sauce. (sorry again, I can't resist)

Re: Sellsiusrealestate.com Yes, we love to share.

We hope to be more than a listing/advertising service. Certainly members can list & advertise any type of property on the site. Those who are involved in real estate need & want more than that, we think. They need & want a place to exchange information, ask questions, learn & help others, & themselves, be better at what they do. They want to go to a place they can trust to be fair with them & give them value. There are any number of advertising & listing sites & if people could afford it, they would use every one to reach as many potential buyers as possible. We believe they can use our website as well to help them. We will use our brand to promote our members, their business and their listings.

We are NOT an advertising based model like zillow--we are member subscription based. We prefer to NOT use an advertising model becasue we think it might make us beholden to advertisers & influence how things are done to meet their demands, wants & needs. (This is not to imply that's what all advertising based sites are doing) We want to answer only to our members & please them. We are also going grassroots because we frankly don't have the funding of zillow. We have been approached by investors but have declined because of concessions which we believe would compromise our commitment to our members.

Only time will tell if consumers & real estate professionals find the sellsius real estate community valuable.

Thanks,
Joseph Ferrara

As a real estate professional, I find Zillow always entertaining, often helpful, but never the final word -- and most of my colleagues, and most home owners and buyers in this area have a similar opinion.

Entertaining: I can only wish that our local MLS or my own broker's web site -- or for that matter a competing broker's site -- provided the same kind of intuitive, information-rich map-based information.

Helpful: Zillow's always the quickest way to get an idea of a property's boundaries, which is particularly helpful in non-homogenous neighborhoods. It's also the quickest map-based way to get nearby recent sales comps. Sure, I could head over to our MLS, import a bunch of data using Homeworks, export it to Excel, correct any obvious address mistakes, replace all city codes with their real names, then export it from Excel to MapPoint...but Zillow is just a whole lot easier. (Actually, in all fairness, our MLS has just released a significantly improved comps feature, which includes mapping capabilities...but it's still not that great.)

Never the final word: Zillow's been around long enough for people to understand that it's simply one of many a starting points -- a slick and convenient one, to be sure, but a starting point nonetheless. Other starting points include my own regression analysis, a quick look at recent comps, maybe an old appraisal if the client has one sitting around. Only after visiting the property in question, looking it over several times, then going back to my office and thinking about it for a while do I commit to a more specific, and hopefully more accurate, number.

Thanks ; I'm glad you find our raw data useful and I will try and address as many of your assumptions & questions about Zestimates as I can.

To start with the assumptions;

Assumption #2 is generally incorrect but can be correct in some edge cases -- remember that the public record is actually being double-checked by many human beings. I've toured the assessor and recorder's processes in Seattle (highly recommended) and there's not much room for an error to make it the whole way through that system. It does happen, but in the vast majority of cases, our inputs are accurate. For those edge cases where they aren't, we provide the "My Zestimator" tool. There are more innovations in the works that will add additional 'double-checking' -- keep watching the site.

The other reason we built "My Zestimator" was so that owners & investors could do "what-if" analysis on the impact of renovations and other improvements. Even if the raw data were 100% perfect, Zillow would have included this tool.

Assumption #1 is incorrect and is explained by my earlier comment -- current market conditions are implied by sales -- it is information that is implicitly included in the most recent sales transactions. The price a house sells for today implies something about its current market. Likewise, the change in prices over time imply something about the changes and the trends in market conditions. This is why recent comps are one of the most valuable inputs to any calculation of house values.

To answer your question about outliers, Zestimates are the output of, and not an input to the algorithm. So, an inaccurate Zestimate can not negatively impact the neighbors' Zestimate. We've found the problem with Mr Oshin's Zestimate and hope to correct it soon -- but until then, it will not impact his neighbors' Zestimates. The Zindex algorithm does use Zestimates as an input and could therefore risk being skewed by inaccuracies -- so we use a "median" Zestimate value to calculate it. As with sold data analysis, a median is the most effective way to exclude outliers (edge cases).

I've given the call for a 'disclaimer' a lot of thought -- frankly, the content on our site already directly addresses every single concern I've heard raised about Zestimate accuracy. I urge you to read it.

We have the necessary fine-print too (see terms of use), but we go further than a typical disclaimer would in our help pages -- and in our reporting of Zestimate accuracy which is highlighted in the search box on the home page. Zillow's biggest critic in this (Greg Swann), now agrees that the necessary caveats are addressed on our site -- Greg's still not a fan of linking to that content -- but that's just the way websites work. Our audience gets it; don't expect any further changes here, but do please read our help content.

If you've visited Zillow, you'll know that we don't let people tell us what their Zestimate should be or contact us to remove Zestimates from the site -- I hope it is obvious that this is critical to the integrity of the Zestimate service and its usefulness to our audience. Future features may address the spirit of what you are suggesting here but expect us to come up with more innovative solutions in that regard.

I personally think comparing Zestimates to CMA's are an apples to oranges issue. IMO, some people miss the most important distinction between the two; a Zestimate is a mathematical calculation of what a house could possibly sell for today whereas a CMA that includes a listing price recommendation is a prediction of what a house will sell for at some time in the future (after the house has been staged, listed and marketed). So, CMA's need to forecast value -- Zestimates do not. I have extreme respect for anyone who can do that well in a changing market. Zestimates and CMA's should start to approach each other in a static market but when last did we have one of those? Note that the time between a house selling and Zillow receiving that transaction record is significantly shorter than average sale lead times, so producing an accurate CMA is more of a challenge (IMO).

I think that addresses all of the points you've raised. More than healthy debate (which I also enjoy), these discussions help clear up confusions about Zillow and give us in Seattle lots to think about as we prioritize next features. Likewise, I hope that by getting into some of the details here, we're arming your audience with the knowledge they need to effectively use Zillow as a tool in their discussions with clients.

Thanks again,
David

PS. Question RE: Sellsius.com; what's the latest scoop on the launch of your listing advertising service? Anything you can share?

Ding ding ding!! This one's over folks. Tell the trauma team to get ready for a delivery.

To answer you question, I have no expectation of a zestimate's accuracy. It may be accurate or grossly inaccurate. I am unable to KNOW just by looking at it. I need an experienced professional to tell me if it's way off or not. If it is badly flawed it is of NO VALUE, even as a starting point, true? So I can't say it's a good starting point before the fact of actually knowing its accuracy, can I? The data on the other hand I absolutely DO NEED. And for that I thank zillow.

So that you understand, our posts simply say that an experienced professional who examines your home (to uncover your unzillowables) & knows the current market & market pool (other unzillowables) will get you CLOSER to pricing your home than zillow. And if something gets you closer then it is a BETTER starting point. It's really that simple.

But while I have your ear let me elaborate a few points. Let me start with 2 assumptions I hope we both can agree on.

1. All other things being equal, pricing a home for sale or purchase relies more on current market condtions and the current market pool than it does on intrinsic value zillow may estimate. Zillow, you must concede, does not factor in current market conditions or the current market pool into the zestimate. Yet this info is valuable to pricing (in fact, zillows data may lag the market because of time delay in public sales data availability)


2. To be valuable, data must be fresh, accurate and complete. If you agree on this, the next question is HOW can you confirm the accuracy, freshness and completeness of data before you input it into the calculations? We say you can't unless a human being double checks it. Freshness, accuracy & completeness cannot be implied, can it?

To this point, you have a Zestimator tool. THe reason we see for the tool is that you recognize that the data MAY NOT BE fresh, accurate or complete & needs to be adjusted. But any such adjustment is not reflected in a newly published zestimate is it? Zillow says the zestimator result is not disclosed, thereby leaving a flawed zestimate on the public table.

Does greatly flawed zestimates in the database affect other properties?

I cite the case of Mr. Steve Olshins in the August comments on zillow blog. If his zestimate is left uncorrected, does it skew other resultst?
http://www.zillowblog.com/zillow_blog/2006/08/e...

I read that your house value reports consider not only actual sales but also unsold homes & I believe I read in Seattle PI that zillow said this made the report more accurate. How can this be true if there may be flawed zestimates of unknown magnitude, since the extent of a zestimate's error is not known until AFTER a home sells. Yet these reports are serving as the basis for interpreting the state of certain markets.

To your statement that unzillowables ARE implied---the proof is otherwise. The zestimator tool exists to adjust the published zestimate.

The existence of a zestimator seems to be an admssion that unzillowables are NOT calculated OR IMPLIED in MY zestimate. Zillow advises using the tool to factor the "ooh and ahh" factors (what we include as unzillowables). If the unzillowables are impled, why would you need the zestimator tool? The oohs and aahs still need to be factored in to adjust the zestimate.
What you said was that unzillowables are impled in the sold comps---but this is different than the unzillowables for an unsold home which is what we are referring to.

Now, even though you may say that zillow discloses its flaws (and we ALL have our flaws) I reply yes, somewhat, but not where it counts most---conspicuously on the homepage & beside the zestimates.

Why does this matter?. Well for one thing, brokers who prepare CMAs are required in some states (PA, NJ & MD) to CONSPICUOUSLY post a disclaimer/disclosure on the CMA to the effect that the estimate is NOT an appraisal & the analysis was not done in accordance with uniform standards of accepted professional appraisal practice. Why are brokers held to a higher standard than Zillow for their estimates? Seems somewhat unfair to favor zillow in this respect, does it not? Have you been asked to put this disclosure in those places? WIll you please? If you did, you would be doing a great service to everyone. We urge you to consider it.


2 questions David: Can an owner opt-out of publication of their zestimate?
Can an owner appeal their zestimate & have it publicly changed?

Obvioulsy public data cannot be prevented from publication. But the zestimate is a proprietary creation using some secret sauce, as you say. There is a difference between the public data & the created zestimate correct? The reason I ask is because as a seller I may not want your value tag on my house, especially if I am selling above the zestimate. It is after all only zillow's opinion is it not?

If a seller asks that you not publish his zestimate and you refuse, what does that say to the market you exist to serve? Would you then be favoring buyers at the expense of sellers? These we believe are important questions.

In the case of Steve Olshin, it seems to us, zillow must either correct his zestimate or take it down. It's equivalent perhaps to an appeal to the assessor to change their opinion on value. Will zillow permit this appeal by an owner? If zillow changes it for Steve, should it not change it for another owner who petitions?

We only ask the questions & state our opinions. Readers will decide if our questions have value or not. They will also decide to ignore our opinions or not. They may ask their own questions. This is all good .

What others are saying-----

Finally, you do realize that the post you are commenting on now was from Inman News, a respected name in this industry.

In addition, the article excerpt was posted on Jonathan Miller's Soapbox. As you know, Jonathan Miller is a respected nationally recognized authority. He is also an appraiser. If you read his comment on this post, it is telling. He says , and I quote:

"In other words, appraisers are needed to bridge the gap between raw data and subtle nuances within a property or market that are not picked up in the data."

This post from Inman would seem to support our position that there exist unzillowables that are not being factored into an owners zestimate.

We thank you for your comment in any case. It spurs healthy debate.

Hi, it's David from Zillow.com.

I've followed your posts on this topic and think I still miss your point -- what is your expectation of Zestimate accuracy?

Most internet users understand that google is not "the oracle" -- and that their search results are just the starting point to finding the answer to their query. Similarly, most Zillow users understand that the Zestimate is a starting point in researching the value of a house -- like google, Zillow just happens to provide possibly the best starting point available online (IMO). Everyone must start somewhere, right?

To quote many of the Real Estate bloggers I read -- "the actual dollar value of a house on any one day is only that value that a buyer is willing to pay and a seller willing to accept". It's not the CMA, it's not the Zestimate, it's not an appraisal and it's not a tax assessed value. That doesn't mean those data points aren't useful.

Further, the posts I've read on this topic fail to show an understanding of how AVM's actually work; the piece that's missing from this commentary is that an advanced algorithm will consider _both_ explicit _and_ implicit characteristics of a house.

Without spilling the secret sauce, here's what I mean:

First, data recorded by the tax assessor (which btw, includes "views"), is explicitly factored into an AVM. Next, data not recorded by the assessor _is_implicitly_included_ in AVM results through the inclusion of past transactions and sometimes, tax assessments. This is how Zillow _does_ account for factors like proximity to schools and arterials. So, the factors you consider "unzillowable" are actually _implied_ by the values of past sales transactions -- of the house in question, its neighbors and its comps. So, when we say that Zestimates get more accurate with time (and transactions), this is often what we're referring to. Does it make sense?

We have a lot of challenges ahead at Zillow, but this isn't really one of them. At least it's not for the majority of visitors who have realistic expectations of Zestimate accuracy. What continues to be our primary challenge is raw data -- to feed both the explicit and the implicit models -- we're doing well there but still have a way to go -- and have a few plans in the works for improving coverage further. Over time, our growing archive of historical transactions will continue to increase Zestimate accuracy as they factor _in_ more and more of the stuff you refer to as "unzillowables".

I hope that sheds some useful light on this topic.