Overstock.com lost a chunk of change last quarter. On Friday, the company reported a GAAP operating loss of $11.2 million and a GAAP net loss of $14.2 million in Q3 2005. So where did they pin the blame for poor results? It wasn’t overly aggressive pricing, too many coupons or escalating shipping costs. According to the CEO, it was an inadequately tested (and prematurely launched) ERP software installation that cooked their goose. That’s right … ERP software …

There’s something oddly fascinating about watching an unconventional company twist over a poorly implemented ERP software roll out. What do you say to the shareholders to cover for the shortfall?

Overstock’s Patrick Byrne starts out by apologizing in Japanese.

Then he regresses into slang. Well, at least he didn’t follow up with “we were SOL after our ERP software went AWOL …”

From the horse’s mouth:

Dear Owners,

Gomen nasai.

Q3 was rough. My bad. I bit off more technology projects than my colleagues could chew. The last bite, an ERP implementation, was one bite too many, and we choked on it. This mistake resulted in two kinds of unfortunate results:

1. Unexpected bad things happened. We tested our new ERP system as much as time allowed, but in mid-August I cut-over earlier than I should have. We anticipated that a small number of background processes would fail, but we had more trouble than expected. Rather than pass the problem to the customer in the form of bad service, we ate the cost by doing a great many things manually that would normally be done automatically. Of course, customer-facing issues (shipping products, processing credit cards, and handling returns) trumped internal issues. One internal issue was our inability to post new products: this gradually reduced the number of products on our site (especially best- sellers). This weakened sales (though we offset this for some time by running $1 shipping). When we were able to upload new products in mid-September, sales snapped back to previous levels.

2. Expected good things did not happen. Some IT projects that were supposed to lift margins through improved efficiencies in logistics or customer service were either late or remain incomplete: some of these projects have since been delivered, some (e.g., a professional customer service application) have been scaled down and partially delivered, and some have been shifted to Q1. Other IT projects that were supposed to yield marketing gains went unfinished. One example: Project Propeller’s test results are extremely encouraging, but only rarely has it been turned on because the A/B test system required to fine tune the system has been low resolution and inconsistent. These projects did not get delivered because the work required to address the ERP problems eventually monopolized development resources and crowded out all other progress (until quite recently).

While I will give more detail on our call, the story behind our results is this: I put off doing things I wanted to do to pursue things that we had to do to prepare for the future; the implementations (and their expected benefits) came in late and rough; and stumbling on the last and most important one set us back a country mile.

Humbly submitted,

Patrick Byrne

Oh, but I kid Overstock. But maybe they shouldn’t have shopped for ERP software in the closeout bin. (Just kidding!)

I like the company. Really, I do.

But I hope everything’s completely worked out with their ERP software by the time I start my Christmas shopping.