Wednesday, May 30, 2007

Real Estate Inventory - How High Will It Go?



February 23rd, 2007: 1,609,000 Single Family Homes
May 30th, 2007: 1,924,110 Single Family Homes
Inventory Increase: 315,000 Houses
or about 100,000 per month
or 1,400 houses every day


I am overwhelmed by the continuing climb of real estate inventory. For people looking for evidence of the tail end of a housing bubble market, this should provide definitive evidence.

In late February when we started tracking the market using Hardtack, we were at 1.6 Million homes listed in the MLS. Today we at 1.9 Million homes, with the rate of growth continuing from it's inflection point March 21st. In early May it looked for a few days as if the inventory was starting to flatten out, but then arced back upwards and resumed its previous slope.

Why is this happening? Firstly, the active season for buying and selling homes is from the spring to mid summer. Part of this is so that families can re-locate and get their children in the new school district before the start of the new school year. In addition to this normal volume, there seems to be a large number of investor and speculators who are trying to sell their investment properties before the market goes any lower, or they run out of money to cover their payments are are forced to join the swelling numbers of foreclosures.

How high will it go? I suspect we will break 2 Million single family homes on the MLS inventory by the end of June, if not sooner. How the market will work through this much inventory is the big, important question for our economy.

Boomerang History - Project Redstone

When we first started our efforts under Boomerang, we created an example for one of our early customers. This customer runs a series of very successful resort, casino and other hospitality destinations across the United States. They ran components of their IT on IBM iSeries mainframes, Java components connected with message oriented middleware, and a large data warehouse. In spite of the hard work of a good sized team in house, they were still unable to come up with solutions to some of the businesses' needs. In this case they wanted to better track and manage how their properties in the lucrative Las Vegas market were performing.

The data that represented what they wanted to know was scattered across several systems, most of which had no normal means of integration, and thus their problem. Using the current industry state of the art would have resulted in another J2EE system that the estimated would take over 12 months to complete.

We applied the principles of "Content Based Integration" to the problem and the Boomerang Feedkit. Our Feedlet's (small Java programs that produce tagged content from enterprise data) scavenged from the Mainframes, the data warehouse and other sources to create a comprehensive, up to the minute picture of several key performance indicators for each property. Furthermore because we created a conforming XML feed for each property, we could use a Fuselet (a small Java program that fuses content based on tag and aspect aggregation) that would create a view of this information across the entire Las Vegas market.

This effort took weeks rather than months, left the important data and systems untouched and in place, and created a new integration channel "out of band" with minimal impact on existing operations. Information on key business concepts, such as how many people were at a property, and what games they were playing (and how much was being wagered) was now available via a simple machine to machine interface to other systems.

Like the original "Redstone" it was a bit of an experiment that was used to prove a point. Our concepts worked, and Redstone was our first successful feed system. Suffice to say the customer was impressed.

Tuesday, May 29, 2007

What is Content Based Integration?








Boomerang is a system for enabling enterprise use of "Content Based Integration". Content is the name for what makes the World Wide Web tick - information in human and machine readable form. For the most part one can think of content as a coherent set of written or illustrative information on a specific subject. When you look at a web site like Amazon.com, you see a wide set of small frames of information - each one of them have words and pictures in them - each one of those frames holds content. Content conveys information along with tags and attributes that describe what it represents and how things relate.

Many IT experts suspect the next big wave of revolution in systems will come from the ability to allow users to subscribe to content on subjects that matter to their work or their interests, and use dynamic web based and desktop systems to assemble this content for them. Think of it as a newspaper that re-publishes itself several times an hour, but only contains the things you want, and are interested in. Another term for this is "Mashups", a term that I think strips some of the meaning of what really should be going on.

You can see early efforts to master this new techniques in products such as Yahoo Pipes, IBM QEDWIKI, and others. The persistent theme behind these examples is that they try very hard to piece something useful and relevant together from whatever they can scavenge from the broader internet. For these current first generation of tools all of the engineering focus went into how to build a wizzy, glittering front end that contained all of the fancy, next-gen Web 2.0 (TM) features on full display.

Sadly for them, once you build these wonderful front end manipulation tools, you are left to discover you don't have much to put in there. Let's face it, there are only so many ways you can glue Youtube, Google Maps and Amazon together. Moreover there are very few (if any) corporations that will find this kind of Mashup compelling enough to spend money on.

The real key is - how do you build a platform that can quickly and easily go from mountains of enterprise data into a steady, rich stream of enterprise content. Content that is formed and tagged, ready for infusion into these emerging Mashup tools.

Hardtack as a Boomerang Application

Many visitors to this site have also taken a look at Hardtack. It's a fun application, but it is also a grand example of how this next generation system works. Data from MLS systems are fed into the Boomerang Feedkit, and transformed into content. That content is tagged based on what zip code it covers, what kind of product it represents, what price point it falls into and when it was sampled. Collections of related zip code content can be fused into towns, towns into cities, then into states and on up to the national level. Product types can be sorted out, price points can be filtered and the content can be aligned and aggregated any way there is a need for it.

When you access the web site, you are really taking different views on the same multi-dimensional set of content data, though you would probably never know unless you read this article.

In our plans for later this year - to allow access to the content in a raw form - either via RSS or XML. This will allow more folks new ways to pipe the content we are keeping to ourselves for now into these next generation Mashup systems. I can't wait to see the results.

Monday, May 28, 2007

Hardtack Birthday Data - 1 Year Trends and Comparisons

I spent the morning creating spreadsheet compares data from Hardtack a year ago and today. There are some interesting trends and information revealed, which I will probably write about more in the future on various weblogs. Anyhow, on the the juicy bits.

Across the 40 metro areas we have year over year data, the trend is: Prices down, inventory up. In some cases way up. Some leaders for inventory accumulation are:

Single Family
Salt Lake City, UT +139.6%
Norfolk, VA +102.4%
Alberquerque,NM +86.2%
Portland, OR +78.6%
Wichita, KS +78.6%
Wilmington, NC +78.6%

Condos
Boise, ID +261.5%
Grand Rapids, MI +212.1%
Portland, OR +162.7%
Alberquerque,NM +109.8%

As far as price is concerned, the trend is down in general, but some markets are showing appreciation. Keep in mind, Hardtack only tracks the seller's asking price. So these do not represent what the properties sell for.

Down Markets - Single Family Median Price
Salt Lake City, UT -15.2%
Bakersfield, CA -13.8%
Boston, MA -11.7%
Miami, FL -10.4%
San Diego, CA -10.2%
Orlando, FL -9.9%
Sacramento, CA -9.6%
Las Vegas, NV -9.2%
Cleveland, OH -9.1%

Up Markets - Single Family Median Price
Austin, TX +20.0%
Raliegh, NC +14.9%
Bloomington, IL +10.3%
Buffalo, NY +8.3%
Syracuse, NY +6.7%
Madison, WI +4.2%

Right now to look across the 40 metro area condo market is to glimpse insanity. The apparent overbuilding and speculation that had taken place (and in some markets may still be taking place) is impossible to ignore. Many people are going to be damaged or ruined financially from this.

Down Markets - Condo Median Price
Salt Lake City, UT -30.8%
Tampa, FL -18.3%
Syracuse, NY -14.6%
Phoenix, AZ -12.2%
Boston, MA -10.8%
Birmingham, AL -9.2%
Richmond, VA -8.2%
Miami, FL -8.1%
Bakersfield, CA -7.5%

Up Markets - Condo Median Price
Cincinnati, OH +40.8%
Austin, TX +27.2%
Omaha, NE 24.8%
Oklahoma City, OK 22.4%
Raliegh, NC 21.6%
Memphis, TN +20.6%
Casper, WY +15.7%
Boise, ID +15.5%

Lastly, and most curiously, it seems there is a bit of a land rush in many midwestern areas. After some discussion with family in the midwest, it would seem that there is a push to snap up tillable land to put into corn production. In essence to grow more ethanol. There is something of an ethanol mania sweeping farm country, with many farmers seeing this as a way to make a better living and help ensure United States energy independence at the same time.

Good review of Hardtack

Jay, The Phoenix Real Estate Guy, has given Hardtack a very good review and named Hardtack his Cool Web Site of the Week. Of interest in his review is the comparison of numbers produced by Hardtack, Realtor.com, Trulia, Zillow, and ZipRealty.

Project Hardtack - One Year Anniversary

One year ago, on Memorial Day weekend, it was finally time to put some of the ideas we had selling to another test. Since early April of that year, I had had the notion of using Boomerang's pattern of scavenging data, refining and distilling it into information, and directly publishing it to tackle a dark corner of the economy: Real Estate metrics.

True you see a lot of information published about real estate, some of it comes out every month. I was unhappy because it had been processed and filtered and "seasonally adjusted". In essence I was getting the feeling that the information was so highly processed that the numbers were saying whatever the real estate business interests wanted it to say.

I decided to build the first version of Hardtack - something I called "Ivy Mike". Named because it was a successful fusion device - the first actually.

Ivy Mike was largely a very primitive experiment with some early Boomerang concepts and code, but as it ran it certainly was able to scavenge and collect the housing data I was looking for, and over the course of the summer of 2006 revealed the early weakening of the housing market and the deflation of the bubble.

In January of this year we decided to move ahead with a full scale system that expanded from Ivy's 40 large metro areas to Hardtack's 300 or so. That effort led by Rob and Kevin is the web site that one year later is beginning to get broader notice.

Hardtack has been a great demo platform for Boomerang, and an excellent vehicle to refine and enhance our platform for hosting Feed technology. My thanks to everyone who has worked to make it reality, or who have given it a look over and learned something new.

More recognition for Hardtack

The folks running the Maine Bubble Blog recently found Hardtack and posted about it. Hopefully they find the site useful and eye-opening.

Saturday, May 26, 2007

San Francisco feeling the pain

Folks up in San Francisco are well-aware of the inventory surplus hitting their city.The San Francisco Real Estate Blog also posted a blurb about Hardtack.

Friday, May 25, 2007

Nice write-up at BusinessWeek

Dean Foust provided a nice write-up on Hardtack and OSG, Inc. today in BusinessWeek's Hot Property blog. The exposure is much appreciated and has given us a nice boost of traffic.

Existing Home Sales - What Hardtrack Is Telling Us

The National Association of Realtors published figures for existing home sales for April today.

Existing home sales fell in April

The median price of a home fell to $220,900, an 0.8 percent fall from the midpoint selling price a year ago. It marked the ninth straight decline in the median price.

Comparing the national sales median price, $220,900. Checking the median price from Hardtack, it reports the median number as $276,002. Why the difference? Let's dive into the guts of Hardtack and figure out.

Hardtack works by reading data from the web front end of the MLS systems. This means we are tracking the prices that people are asking for their properties, what they hope to get. Nation wide, in almost every market the inventory keeps climbing higher, over 1.9 Million single family homes on the market as of today. In many places across the US, sellers are asking more than what most buyers can now afford to pay. Either because they think their property is worth more than the market will currently bear, or sadly because they have used modern exotic financing approaches such as cash out re-fi's and home equity lines of credit and they are trying to sell for how much they owe, which in some cases is more than they can possibly get.

Median Selling Price: $220,900
Median Asking Price: $276,002

That $56,000 gap signals a fundamental sickness in the housing market, one that will need to be worked out over the next few years.

Thursday, May 24, 2007

Nice Plug for Hardtack

Hardtack got a nice plug today from SeekingAlpha. In today's article Hardtack was listing as the "Web Site of the Day," which generated us a nice bit of traffic.

Friday, May 18, 2007

Hardtack - Real Estate Inventory Tracking

http://hardtack.osgcorp.com/

Hardtack provides a view into the real-estate inventory available at various levels of the "real estate hierarchy": nationally, regionally, by state, MSA, and at the single-city level. The ability to see the amount of inventory available and the trends in inventory availability has been, before Hardtack, unavailable to the people who need the data the most: home buyers and home sellers.

The Hardtack application demonstrates two of Boomerang's key strengths: the power of the feedlet idea and our ability to create snazzy UIs.

The backend application of Hardtack consists of feedlets that produce a variety of content from some legacy data sources. Whether the data sources are legacy applications, databases, other web resources, flat files, or anything else, feedlets provide a layer of abstraction that expose the data source to any XML-capable system in the environment.

The UI of Hardtack fuses together a variety of data and displays it multiple ways: maps, charts, and in tabular format. The result is (hopefully) a very simple, compelling view into a particular dataset.

Take a look at Hardtack and let us know what you think. Your feedback helps us drive new features!
http://hardtack.osgcorp.com/.

Thursday, May 17, 2007

Boomerang - a New Approach to Business Intelligence

Businesses large and small have been investing in IT middleware and integration solutions for nearly 20 years. The IT stack that has accumulated is often bloated and brittle; layers upon layers of technologies that, while well-intentioned, have paralyzed the business and left it feeling like a mammoth at the La Brea Tar Pits.

A by-product of this bloated stack of MoM, EAI, J2EE, Corba, XML, SOA, and .Net (just to name a few) is that often times the business side of a company loses its ability to change and adapt quickly to new business prospects. Why? The applications and technology necessary for the business aren't agile enough to support the business needs of new applications, new visualizations into customer data, etc. We've seen a recurring pattern amongst our customers: the business sees potential in a new area and requests IT projects to support this new venture, IT responds with a multi-million dollar request and a 6- to 12-month project schedule. With all the advances in technology and the countless number of intelligent software engineers employed by these IT departments, shouldn't projects cost less and take a smaller amount of time?

Boomerang is a new approach to building Business Intelligence applications. The way we build applications is different: less bloat, less complexity, more value. The timeframe and costs associated with our projects are different as well; projects usually span between 1 and 3 months and the costs pale in comparison to the high software licensing costs of current BI and middleware vendor packages.

We aggressively simplify the technical and management aspects to our customers' projects and deliver high-quality, value-added applications.