Apple’s Siri on the iPhone amazes us all with her witty responses and incredible knowledge.  IBM’s Watson Super Computer beat the best brains in the world at Jeopardy and is now tackling real world problems including medical diagnosis and underwriting.  What do these machines have in common?  They both learn from their mistakes.  The promise of Artificial Intelligence or Machine Learning is finally making its way to reality.


As we all reflect on our accomplishments and challenges from this year, how can we advance our companies and learn like a machine?  How can we turn failure to our advantage and fail forward?  Below are 6 ways to start learning like a machine and some examples of agile improvement from the Hard Money Lending business:


  1. Create Goals/Objectives/Tactics: I learned the infamous Goals/Objectives/Tactics approach to business strategy during my days at IBM and those disciplines and metrics serve as the basis for machine learning.  The key is to establish metrics and measure.  Quantify everything.  Once the data is available then bubble up the important which, when measured and attained, will increase revenue and profits.  An example in the lending business is setting a goal to increase the average loan amount by $100k by the end of 2014.  The objective might be to increase loan amounts from $250k to $350k and the Tactic to accomplish the objective might be to acquire better leads by joining a networking group in a higher priced neighborhood.
  2. Acknowledge Failure and Take Responsibility: Failing means you are in the game.  Take solace that the mistake happened live and learn like a machine.  Remember, failure is not fatal, but failure to change may be.  I had a loan that went into foreclosure and I later realized that the borrower’s income was too dependent on the rental income from the property.  Now we apply new and improved ability to pay rules to our underwriting.
  3. Check the Competition and External Data Sources:  When establishing metrics for your business, check out the competition and industry sources for clues.  For example if a lender is measuring and trying to increase their interest rates and margin, check out rate sheets from other lenders.
  4. Garner Facts from Relevant Data and Induce Change from Data Nuances: Facts are that the new Dodd Frank law is changing the lending landscape.  Smart lenders are looking at these new regulations and are crafting new loan programs that fit into new niches created by these new rules.
  5. Read between the lines and Recognize Patterns: Sometimes we have to learn the hard way when we have to pay for a mistake.  On one loan we made, a triplex really wasn’t a tri-plex.  The property was a duplex with a bootlegged third unit.  I learned my lesson.  Now I always check building permits on a property at the City’s Planning Department.
  6. Establish new Metrics and Measure Again: Be an “Agile” learner.  Crystalize the important data and discard superfluous noise.  Establish the metric, measure, iterate and measure again.  Repeat this process and improve across your key performance indicators.


Mortgage Vintage, Inc. measured and in many cases attained our 2013 objectives.  Where we fell short, we will iterate and develop clear and concise goals, objectives and tactics for 2014.     We hope your business continues to grow, develop and learn like a machine.  What have you learned this year?