It seemed like such a good idea at the time. Quicken Loans affiliate Title Source had signed an agreement with start-up HouseCanary to build specialty software. The relationship deteriorated, and Title Source sued to avoid the $5 million fee it had agreed to pay. To underscore its determination, it served the lawsuit personally on HouseCanary executives at their trade show booth, right in front of their customers.

Commercial litigation often starts with this sort of bravado, and Title Source apparently felt it had been seriously wronged. But less than two years later, in March 2018, a jury disagreed. It decided that Title Source had used HouseCanary’s secret data to build a competing product and slapped it with a verdict for $706 million.

The back story of this turnabout case provides some important lessons, grounded in the special nature of trade secret disputes. But first let’s understand how a hopeful software development deal turned so ugly.

Real estate appraisals traditionally rely on a physical inspection by a trained professional. But much of the data they collect about a house is already available online, or can be extrapolated. So just like many other industries, this one is undergoing disruption through automation, as firms develop Automated Valuation Models, or AVMs.

Title Source, one of the country’s largest title insurance and real estate valuation companies, signed a contract with software start-up HouseCanary to develop an AVM. The product would be licensed to Title Source for an annual fee of $5 million, but HouseCanary would maintain ownership and rights in its own data. The relationship began to break down over requested changes, and Title Source decided that it wasn’t getting what it bargained for. In the meantime, HouseCanary had shared its secret information with Title Source engineers.

Those same engineers were tasked with building their own AVM. From the perspective of Title Source, this project was necessary in order to cover for HouseCanary’s failures and to mitigate the damage it was suffering. When the time came to pay the $5 million fee, Title Source decided that an aggressive litigation strategy would best serve its needs and reflect its frustration.

This is where the drama begins its teaching. Title Source believed its own narrative, in which it was a victim of HouseCanary’s breach. Its commitment to that narrative masked the possibility that hidden in the records of the two companies’ dealings lurked a massive potential liability. When emails were produced as part of the litigation, it appeared that the engineers tasked with building Title Source’s AVM had used HouseCanary’s confidential information to do it. (One email exhorted colleagues to “think big and wide about how to maximize the value of the HouseCanary data to our business.”)

Some of this evidence was ambiguous, and Title Source argued that what it did was within its rights. But it was looking through its own lens, and didn’t fully appreciate the risk that HouseCanary’s alternative narrative of deliberate misappropriation would be accepted.

Why didn’t Title Source see the potential disaster when deciding whether to sue? The answer almost certainly lies in the emotional content of disputes where information has been shared between companies. The relationship starts, as it must, with declarations of trust on both sides. So when things start to go downhill, disappointment morphs into loathing and a sense of victimhood. Each side, anxious to see its own behavior as fully justified, develops a committed perspective.

Here, the miscalculation had very serious consequences. HouseCanary, having convinced the jury that its information had been misused, was able to present a simple and compelling damage model. It argued – again, using its adversary’s own records – that Title Source planned to save a certain amount of money on each AVM that it sold. Assuming that it would have taken two years to develop that product legitimately, and using internal projections of likely use, the benefit to Title Source worked out to be over $201 million.

The jury accepted that calculation, and added HouseCanary’s lost profits of almost $34 million. Then, because it had embraced HouseCanary’s narrative of a deliberate misappropriation, the damages were tripled, for a total of $706 million.

The first lesson we can extract from this story is cautionary: assessment of risk in potential litigation requires a sober, objective analysis of the evidence, with plenty of “devil’s advocate” scenario planning. This means resisting the kind of confirmation bias that can interfere with management’s ability to appreciate how the same facts can be interpreted in different ways.

Second, and closely related to the first point, we need to be aware that trade secret cases often spring from facts learned in unrelated litigation. In one case I worked on, a patent infringement claim led to the production of records showing that a misappropriation had occurred more than fifteen years earlier. Because the statute of limitations runs from the time of discovery, the accumulated damage claim was massive, and immediately the trade secret claim eclipsed the patent dispute.

Third, the HouseCanary verdict provides a reminder of the flexible and often generous way that damages are calculated in trade secret cases. The aggrieved party can collect not only for the damage it has suffered, but also can recover the perceived benefit received by the defendant, even if it hasn’t produced any profits or even any products.

Finally, this case suggests that companies engaged in collaborative development of technology need to manage the process very carefully to reduce the risk of later disputes. Clear protocols and careful record keeping around the receipt and use of confidential information will go a long way to prevent problems.

And a bit of training in how to draft emails wouldn’t hurt.