Leveraging technology to compete with AVMs — but not in the way you might think

Matt Barr , Communications Director, a la mode , inc.

There's never been any doubt where a la mode stands on Automated Valuation Models (AVMs). Time after time after time after time we've made the public case — often alone, which is odd, considering the bevy of appraisal trade groups — that appraisers are far superior to AVMs, and not only that, AVMs simply don't work on an individual property level.

Want proof? Let's say you could pay $20 to $50 and thereby cut your work on an individual appraisal by about 75 percent. Yet you could still charge $350 (or the going rate in your area) for that 1004. You could get four full–fee reports done in the time it now takes you to do one. Your margins would go through the roof.

What's the secret to this ultimate business success scheme? Simply run an AVM on whatever property you've been assigned to appraise. You'll get a dead–on accurate opinion of value, the very best comps with bulletproof adjustments, and your work involves eyeballing the data, filling out the form and taking the pictures.

Except you don't. Only nine percent of appraisers in the 2007 October Research National Appraisal Survey said they use AVMs as part of their valuation process. Of that nine percent, 86 percent say that less than a quarter of their work involves the use of AVMs.

Appraisers want to make money as much as anybody does. Nine out of 10 of you aren't refusing to use AVMs on principle. They wouldn't save you work because they don't work.

It's not your problem (or ours) that AVMs don't work, but it is your (and our) problem that lenders are interested in them and are rooting for them to hurry up and work (and have been for 10 years). When cast this way, competing with AVMs becomes a marketing challenge more than anything else. You'll get the business an AVM might have gotten if you convince the decision maker that the extra time and (borrower's) money are worth it. They are: Your work is more accurate, more supportable, more transparent, accompanied by far superior customer service and backed by your professional reputation as well as your E&O insurance.

But there are technology issues involved as well. AVMs are not (remotely) as reliable as you are, but they are faster and cheaper. You don't have to improve on your reliability to compete, but improving on your turn times and fee only makes the decision to use you that much easier.

A subset of competing on price is that, naturally, you need to profit from your work. Slashing fees while doing the same thing with the same effort in the same amount of time doesn't cut it. (Nor should it.) So how can you do more work, with less effort, in less time? How in other words can you trim your turn times?

If you use formfilling software, a digital camera, a PDF printer, a database of comps, and any of dozens of other relatively recent innovations you already have reduced your effort and increased your speed considerably. When someone grumbles that appraisers need to adapt to a changing world where things happen so much faster, it's because they don't realize that appraisers are early adopters of most technological innovations already. Cell phones, desktop software, mobile devices and applications, digital imaging, electronic document delivery — no other participant in the home selling or mortgage process is as technologically advanced as your average appraiser.

So the labs project represents a great opportunity to speed the next time–saving innovations to market. We're excited about the possibilities. Among them are data sharing and improved field data gathering, which a month or two in to the labs project we're already discussing at length. There are other big ideas such as in the realm of converting paper to digital files, automating compliance, eliminating field keystrokes with voice recognition, and much more that will percolate in the labs in the coming months.

If we build it, and it works, you'll use it. If it doesn't work — like AVMs — you won't. The labs project is where we hash out what works and what doesn't, with input from you.