Decide.com
Decide.com was a price-prediction engine that told consumers whether to "buy now" or "wait" for a price drop or a newer model. Using sophisticated machine learning to track millions of products, it aimed to bring "Wall Street-level data" to the average shopper. Despite high user satisfaction and significant funding, the company was acquired by eBay for its data talent and the consumer-facing service was permanently shuttered.
The Autopsy
| Section | Details |
|---|---|
| Startup Profile | Founders: Oren Etzioni, Hsu Han Liu, and others Funding: ~$17M (Investors: Madrona Venture Group, Maverick Capital, Vulcan Capital) |
| Cause of Death | |
| The Critical Mistake | Underestimating Retailer Resistance: Major retailers (like Amazon and Best Buy) began to realize that Decide.com was hurting their margins by encouraging "waiting." In response, some retailers began implementing "dynamic pricing" that was designed to confuse or circumvent the very algorithms Decide was using. |
| Key Lessons |
|
Deep Dive
In the post-mortem reflections, including the founder's thoughts in "Why We Sold to eBay," the focus was on the transition from a consumer "shield" to a corporate "sword." The Accuracy Standard Decide.com was famous for its "Confidence Score." If they told you to wait, and the price went up, they lost user trust immediately. This meant they had to over-engineer their predictions to be 90%+ accurate, a technical hurdle that most of their competitors (like Honey or CamelCamelCamel) didn't bother with, allowing those competitors to be leaner and faster. The "B2C to B2B" Pivot Before the acquisition, the team realized that their data was actually more valuable to the sellers than to the buyers. Retailers wanted to know when their competitors were going to drop prices. While the founders initially set out to help the "underdog" consumer, the economic reality was that the "big players" in the e-commerce space were willing to pay much more for that data. The Legacy Decide.com is considered a pioneer in the "Transparent E-commerce" movement. Its core technology didn't die; it was integrated into eBay's "Seller Hub," helping millions of small businesses price their goods correctly. The "Wait/Buy" feature they pioneered can now be seen in various forms across travel apps like Hopper and specialized price trackers. The founders and leadership team went on to hold high-level positions at eBay and later Madrona, continuing to influence how machine learning is used in global trade.
Key Lessons
The "Impartiality" Tax: If your business model relies on telling people not to spend money, you will struggle to find partners in the retail industry.
Data is More Valuable than the Interface: For many startups, the "consumer app" is just a demo for the underlying engine. Investors and acquirers often value the "how" (the algorithm) more than the "who" (the end user).
Timing the "New Model" Cycle: Decide was incredibly successful at predicting "Product Refreshes" (e.g., when the next iPad would launch), proving that transparency in hardware cycles is a high-demand consumer need.