Evan Schuman's StorefrontBacktalk

JDA/I2 stressed that the contract used standard language (“this agreement constitutes the entire agreement” and “this agreement supersedes all prior communications between the parties including, but not limited to, communications with i2’s sales representatives.”). In other words, the vendor’s policy is to decidedly not stand behind any claims they make when selling the product/service. If it’s not written into the final contract, it’s not relevant. As one i2 document said: the contract “prohibits reliance on pre-contract communications.” Good to know.

One example of “pre-contract communications” showed up in an i2 sales rep’s notes on a sales meeting. The project’s requirements called for the database’s ability to handle one million items. The problem, as the sales rep noted in his report to his colleagues (but not the customer) was that the database routinely crashed when it hit 300,000 items. Oops.

The vendor also challenged claims of lost profits on the ground that those losses couldn’t be positively and solely attributed to the software problems.

The vendor also argued that it had no obligation to reveal to Dillard’s an SEC Consent Order or complaints from other i2 customers. “I2 owed no duty of disclosure to Dillard’s, as a matter of law.”

Among the reports that the vendor said it had no duty to disclose was a report they had commissioned from Michael Cusumano, an MIT professor. He had been hired in early 2001 “to investigate i2’s product development process,” according to a Dillard’s filing, which elaborated: “As part of his investigation, Cusumano interviewed scores of i2 employees. He then delivered his report to i2 management on June 5, 2001, only six months after Dillard’s signed the license agreement and just as i2 was beginning the software implementation at Dillard’s. Cusumano’s report disclosed the fraudulent practices at i2, including word-for-word quotes from interviews of i2 employees who admitted they were engaging in fraud.”

As for problems with other customers, that was referencing KMart and Best Buy.

“These other retail customers licensed the same products at the same time and experience the same problems—scalability, missing functionality, unsuitability and buggy code—as a result of the same cause: software that was not ready or functional.”

On scalability, Dillard’s said the i2 claims were specific, with i2 having promised that the software “could scale to 15 to 18 million SKU/store combinations.” As for speed, it compared the promised software with the existing software from Manugistics. “While Manugistics’ software would take 42 days to run Dillard’s millions of SKU/store combinations, i2’s software would take only 4 to 5 hours for a complete rerun from scratch and a couple hours for daily updates.”

Actual results, according to Dillard’s filings, were quite different, with one i2 package “able to process only 9 million SKU/store combinations in 40 to 70 hours.”

At one level, this is nothing new. Software sales reps have never been hired for their humanitarianism or their honesty. It’s buggy and doesn’t scale? Yeah, it’s called enterprise software. If it worked perfectly, it would be called a demo.

I2 was right in that a vendor has no legal obligation to point out customer complaints. But what about the other shoe? When caught, you fix and overcompensate. The fact that a JDA unit today deals with accusations of sales lies by pointing to the contract that says “if it’s not written here, it doesn’t matter,” that says quite a bit to current and prospective customers. This is Reason #8956 why public courts are a good thing.