Home » Lean Six Sigma Articles » Oops!

Oops!

Recently, CheapTickets.com made a little error loading rates for flights to Reykjavik, Iceland. Round trip airfare from New York for only $61, a far cry from the $787 it should have been. About 800 people took advantage of the glitch during the 22 hours it was available.

It seems there's a website where frequent flyers share this kind of information: Flyertalk.com. By the time the listing was posted on Flyertalk, it was late evening in Iceland and no one probably caught the glitch until the next morning.

Unlike some carriers who do not honor their mistakes, CheapTickets and Icelandic did honor the fares. It's cheap marketing, because all of the press services picked up on it and spread the word nationwide.

Software Bugs

There are two main kinds of software bugs:

programming bugs-logic errors, mathematical computation errors, etc. These are harder to find and fix because you have to debug the code, fix it, recompile it, release it, and rerun the job.
data bugs-rate table errors. These are easy to find and fix.

Clearly, every company has some kind of ordering and billing system. Rates for products and services are loaded into rate tables that the ordering and billing systems use. Load in the wrong rate and you could lose money because of underpricing like CheapTickets or lost sales because of overpricing.

When I worked in the phone company, there were thousands of rates for phone service and products that varied by city, county and state due to regulatory requirements. All of these rates had to be kept up to date. Sometimes they were, sometimes they weren't.

We used to rate long distance calls by time and distance. I've seen programming bugs like two minutes 30 seconds rounded down to two minutes that cost the company millions of dollars. I've seen rate table errors cost millions as well. I've also seen the kind of negative publicity you can generate if you try to collect the revenue you failed to bill correctly. The press came down on the phone company like a ton of bricks.

Mistake-proofing

How could CheapTickets have mistake-proofed their rate tables? Seems obvious that any international fare under $200-300 should raise a red flag. Similarly, an economy fare within the 48 states should cost under $400. Could a program be developed to analyze table updates as they happen or to analyze the tables for these kinds of anomalies? Sure. Would it be worth it? Certainly, because machines are more precise than people are when it comes to examining data. Of course a programming bug in the analysis program could raise too many red flags or ignore some obvious problems as well.

Have you done everything you can to mistake-proof your rating and billing programs? Have you put similar safeguards in place on the purchasing side of the house?

Internet communities now make it easier than ever for huge numbers of people to take advantage of corporate mistakes in the long minutes before you discover the problem and correct it.

Sure there are ethical arguments about taking advantage of irrationally low fares, but you posted them; therefore they must be valid. It's never the buyer's mistake, only the company's. Will you honor the mistakes you make? If you want to stay on your customer's good side, you'd better honor them. There's no guarantee that an error like CheapTickets will garner the same kind of publicity. You may just have to eat the loss, but it still makes a good story that customers will tell to others.

Haven't you waited long enough to find ways to mistake-proof changes to your financial systems?