Warning Zillow Investors - You Should Give Your Money to Trulia

It was just announced that Zillow received another $30,000,000 in funding to go along with the $57,000,000 they have already received in two previous rounds of investing.  According to the article, the money will be used to accelerate the offerings from Zillow to homeowners and real estate professionals.

I'm not sure what analysis the investors used to determine Zillow was worthy of additional funding but I can tell you they probably didn't talk to any real estate professionals, they probably didn't do a random check of the zillow zestimate to check for accuracy and they didn't compare site traffic growth of zillow to their main competitor, trulia. 

If they had here is what they would found:

As the marketplace leader here in Loudoun County Virginia, a county with a per capita household income level that is 2nd in the entire United States, I have listed and sold close to 200 properties in the last 4 years.  In that time I have consulted with over 500 homeowners on the value of their houses and have been inside well over 2000 homes in that same time period.  I can tell you that without a doubt, zillow's method for valuing a home is a complete joke.  In fact, I don't think they have even tried to fix the algorithms used to create their zestimates since they rolled them out.  Last November I wrote a post comparing the AVM websites and their results.  Zillow was a featured site and they valued the subject property at $482,684.  At the time their estimate was too high by about $65,000 or 15.5%. 

In the last 10 months, the market has gone nowhere but down.  The last sale in my neighborhood of the exact same model as the subject property, with the exact same features, backing to the exact same trees was $389,000 on July 31st, 2007.  Another home on the next street over, same model, same features, sold for $425,000 in April.  There has been a trend downward in the past year and these two sales illustrate that.

So what does Zillow say the house is now worth?  $528,788!!  What kind of algorithm are they using, or what kind of data are they gathering that could INCREASE the value of a townhome $46,104 when in reality the price of the home went down $15,000+?  How is that possible?  And why would someone give $30 Million to a company that is making these ridiculous calculations and holding them out to the general public as a valuation?  It is obvious that they have given up trying to make their zestimates accurate and are focusing their efforts in other areas.  I say this because there is so much inaccuracy in the data (they change the subject property from a townhouse to a single family home) abondonment could be the only explanation. 

But maybe their new features are generating traffic and page views and keeping people on their site longer and the zestimate is just a tool to get them to the site.  Let's see. 

In this graph, it shows that the overall traffic over the last twelve months has gone up a whopping 1.8%.  Compare this to Trulia's traffic growth rate in the last year of 301.6% and it is just a matter of time before Zillow gets passed.  I'm not sure the Zillow investors saw these numbers before they made their $30,000,000 bet but maybe they are looking at how long visitors are staying on the site.  Maybe Zillow is increasing the length of time and page visits of their visitors through superior features.  Let's take a look.

Once again, this graph shows zillows pages/visit has increased 1.1% in the last year while Trulia's has grown by 334%.  Whooops.  Maybe the investors were looking at length of time per visit.  Maybe the information on each page of zillow is so comprehensive they can only absorb 10 pages of information in the time it takes a Trulia visitor to look at 25 pages.  Let's see.

I almost found a statistic that showed zillow as the better investment.  But it turns out that the average stay of a trulia visitor has been longer since April.

In my previous life I was a financial analyst for a mergers and acquisition firm and an investment banking firm.  I can tell you that this is by no means a comprehensive analysis of zillow and it doesn't take into account factors that outsiders are not privy to.  Maybe they have another killer application in the pipeline.  But as a representative of both of their target audiences, homeowners and real estate professionals, I can tell you that my money is on Trulia.

Update:  Joel over at FoREM just published this post about the same topic.

Further Reading: Trulia Leaps Ahead To Become One of the Top 3 Trafic Drivers To Real Estate Companies Nationwide - AOL Money and Finance

Comments

[David G from Zillow.com]

Hi Tony,

As previously mentioned here, Compete's data for Zillow is not accurate. 4,4M users visited Zillow in August.

[Tony Arko]

I was waiting for that David. But I figure, if it is wrong for you, it is wrong for Trulia, too. Which site would you like me to use for comparisons?

[David G from Zillow.com]

Hey Tony,

I can't speak for Trulia's numbers. Unfortunately only measured traffic is accurate. Most companies report on their internal measurements - no doubt that's partially to correct perceptions created by the likes of Alexa and Compete. I totally understand that a relative measure could be good enough for analysis purposes. My concern about Compete is they don't show the 17% growth we've measured y-o-y - which makes me think there's a problem with more than just the volume of their numbers.

[Tony Arko]

David, I am sure that your internal numbers are more accurate than compete's. Unfortunately, we outsiders have nothing else to go on. And I am sure that your investors had more accurate information to go on when they made the $30 Million commitment to zillow. I just hope you guys can come up with something stronger than the "zestimate" for $30,000,000. Something that homeowners and real estate professionals can actually use with some degree of confidence.

[Pete Flint from Trulia.com]

We agree that compete.com's numbers are not accurate, but the trends/growth rates are fairly similar with what we see from our internal numbers. There is often a large (but somewhat consistent) deviation from our internal numbers. The same cannot be said for Alexa whose data does not correlate with our internal data.

It is worth noting that compete's traffic is US only, but internal numbers include everything, which may explain some of the deviations.

We find these services fairly useful for a quick check on relative traffic performance. Hence directionally they are useful, but you don't run your business from them...

For another angle on traffic trends, Trulia.com announced some data this week showing the downstream visitors that we send to real estate franchise and brokerage sites. For a broad selection of partners we now drive on average 11% of their inbound visitors and alongside google and yahoo as a top traffic referrer. http://www.truliablog.com/?p=207

That's a very meaningful % for our advertisers and Industry partners and it's growing!

[Frank Borges LL0SA- Broker FranklyRealty.com]

Congratulations for getting somebody from both Trulia and Zillow to comment on your post.

As for the Page Per Visitor, I don't like that metric. Those numbers can sometimes be way off if a progressive site uses Ajax or just plan ole good design to reduce the number of clicks to deliver the same information for the consumer.

I know Realtor.com is very "NEXT" Next Next. I don't think you can show 100 results at a time, only 10. Why? Maybe in part to keep the pages per visitor number artificially high and to show more ads.

Frank- Broker FranklyRealty.com

[fsbo]

I consider Zillow to be a very good Internet project. They help realtors and FSBO seller and buyers not only to search but to compare and so on…