It seemed time to check on Zillow again and, while we were at it, to take a look at a couple of relative newcomers to the home values genre of websites.

Primarily we were curious about Zillow's Make Me Move feature that allows owners to enter information about their house and a price that would make them take notice even if the house was not actually for sale. Use of this novelty was increasing by leaps and bounds as was another service where sellers or their agents can list, for free, houses that are actually for sale.

Well, the popularity of both continues unabated. When we last visited on December 18, shortly after the features were announced, 9,880 homes were listed for sale by sellers or their agents and there were 5,431 owners daring someone to make them move. Today those numbers have skyrocketed to 59,006 "for sale" entries and 31,451 homeowners putting out feelers.

At about the same time Zillow launched a "wiki." Originally seeded with dozens of well-written staff submissions on buying, selling, building, and remodeling houses as well as information on mortgages, "wiki" now has around a thousand entries submitted by readers. We pulled up several dozen at random (many of the entries are apparently marketing pages for local, national, and international agents) but found many to be empty of everything but a headline.

Zillow now claims to cover 67 million homes, little changed from December.

We had been told about two other real estate oriented sites of relatively recent origin; and A visit to the latter left us with no clue what-so-ever as to what it was about. It may be generational; aimed at the people who flock to Myspace and its peers and we are just too old. But, after investing 30 minutes in reviewing it we moved on. If anyone can provide a brief outline of how to use it and why we should we would be delighted to receive it.

The site, however, seems poised to give Zillow a little competition. It is sponsored by Fidelity National Information Services, a provider of automated valuation services and software for lenders and realtors and claims to cover 100 million property listings. It shares many of the attributes of Zillow but Cyberhomes has not (yet) imitated some of its predecessor's quirkier features.

Cyberhomes has satellite images of the world that can be narrowed down to aerial views of individual homes. The birds-eye views incidentally have improved dramatically on Zillow and both sites now offer pictures of individual houses that are easily identifiable if one is familiar with the property.

Cyberhomes offers a wealth of data on states, towns, local areas (probably Census tract based.) Educational information is provided on availability of public, private, and charter schools, expenditures per student, student/teacher and student/counselor ratios and individual schools are named and enrollment sizes given. Demographic information includes the area's median age, education level, income, employment rates, and gender breakdown. Wonder about the weather? Average temperatures in July and January, rain and snowfalls, and the number of sun-filled and rainy days are given. All of this data is readily available from other sources, but it is helpful to have it in one place.

Like Zillow, Cyberhomes allows visitors to upgrade houses with new kitchens, baths, and amenities such as hot-tubs or decks and see the impact on values. There is also an easy to use comparables function but we found a wrinkle or two there which we will explain later.

What continues to amaze is the difficulty that Zillow has in nailing down data and now, with Cyberhomes to compare against, discrepancies become even more apparent. Zillow may be right, Cyber may be right, but in our very limited experiment it was impossible that they both were much of the time.

We looked at (or at least for) four homes in four different states for which we had firsthand knowledge. We had looked at all of these twice before in evaluating Zillow and found that two of them had dramatic changes in description and values in a very brief period of time. Things have changed again. The properties are located in Brunswick, Georgia; Arlington, Virginia; Salt Lake City, Utah; and Lexington, Massachusetts.

First the respective data:

BR BA SF Price
Arlington, VA Unk 2 1,984 $699,200
SLC, UT 5 2.75 3,126 $503,403
Lexington, MA 5 2 1,112 $1,470,994
Brunswick, GA Unk Unk Unk Unk

BR BA SF Price
Arlington, VA Unk 2 1,984 $709,424
SLC, UT 5 2 3,380 Unk
Lexington, MA 3 2 1,650 $700,132
Brunswick, GA Unk Unk Unk Unk

The most glaring inconsistencies, both between the two sites and with reality occurred with the Lexington property. This is a nice three-bedroom cape in an area of similar homes but where significant tear-downs and rebuilding has been going on. Similar homes in the area are valued in the $500,000 price range so when the $700,000 estimate popped up on Cyberhomes we assumed it had fallen victim to a contractor yet the non-price info still indicated it was a modest cape. Then Zillow's data seemed to confirm the teardown theory - but at twice the price. Yet, aerial photos, with a recent copyright date on both sites show the same little cape that has been there for 50 years.

The SLC house came on line with Zillow a year ago with a value of over $1 million. Three weeks later it had retreated to the high $300,000s. No particular argument with Zillow's Zestimate today, but why, given that assessors data is being used by both sites, does Cyberhomes have no value for the house or for any other in that very upscale neighborhood. And, incidentally, the house has three full baths. The Virginia house was consistent between sites but one wonders why information on the number of bedrooms is missing for a house built 85 years ago.

We keep checking to see if coastal Georgia has joined the world and Cyberhomes indicates that it has the area fully covered. On closer inspection there was information on only 11 homes in a town of 15,000 and the closest of these to our target property, a small ranch on about 20,000 sf. is described as being on 10 acres and, while the address is right, is placed in the wrong county. There are a scattering of For Sale and Make Me Move homes on Zillow but not much has changed since the last time we looked.

The comparables we mentioned earlier don't seem particularly helpful on Cyberhomes. The Lexington house had several dozen "comps" but they seem to have been put forth without rhyme or reason. Some were nearly two years old (appraisers and real estate agents try to limit themselves to six months), were scattered all over town, and ranged from two to five bedrooms with about a 900 sf size spread. In other words, there seemed little basis for comparing them. There were no comparables for the SLC house, possibly because there were no price estimates in the area. All of the Virginia comparables were from 2006 and 2007 and seemed reasonably close in location although the house sizes were all over the place.

It is hard to see, gimmicks aside, how these sites are going to become more useful unless they can pin down their data. Both may be on the right track by inviting homeowners to verify and/or update the information but how many will be willing to make the effort to do so? Maybe Zillow doesn't care. Its business plan appears to involve becoming a link between buyer and seller. No theories yet on where Cyberhomes is headed. Maybe accuracy won't be that important to their business model either.