Launching new products is risky. Some companies don’t even bother—they just make acquisitions and assume every third one or so will work out. Others have highly developed processes for de-risking new product concepts.

Last month we asked whether there is a “right” way to generate new business ideas. This month, we’re looking at the next step: once a company has an idea, how do they move it forward?


Amazon has a much-ballyhooed approach to vetting new product ideas: they write a press release for the product before it exists. If it’s hard to write, it probably shouldn’t move forward.

If writing the press release doesn’t torpedo the idea, the release is shown to internal groups and sometimes a few potential customers for their reaction. While exactly how those reactions are observed is unclear, lukewarm feedback means the press release goes straight to rewrite.

We are pretty sure that this is not the only way Amazon innovates. When the company launched a brick-and-mortar bookstore in Seattle in 2015, for example, the whole process seemed a little cloak-and-dagger. If there was a test press release, it was hard to find.

But floating a trial balloon is a good way to avoid investing in the wrong thing. And Amazon’s press release format, which forces plain speaking and discipline, is an inexpensive way to de-risk new ventures very early in development. The best part: pretty much any size organization can use this approach.


If Amazon’s idea vetting is backwards, its overall R&D approach is anything but. As Azeem Azhar points out, Amazon often develops products that use computational capability that doesn’t yet exist—but it’s coming. This approach allows them to think about innovation in months, not years. And it keeps them from getting hung up on current cost structure.

Presumably innovators have used reliable predictors of change like Moore’s Law to anticipate change for decades. But what else can be anticipated? Here is an impressive round-up of exponential change–everything from the plummeting cost of mapping a genome to the growth of the middle class. Put it to good use, please.


When most companies talk about product experimentation, they mean A|B testing. Booking.com is known for this approach, and there is no question that Booking has an unparalleled competency for optimizing every corner of their platform.

BUT: “If experimentation is only used to ‘finesse the detail’ by A/B testing minor changes, you’re wasting the opportunity,” says Stephen Pavlovich of Conversion. Optimization is valuable, but it’s fiddling at the edges–it’s not driving real greenfield innovation.

How do you ensure that what you are testing is "real" innovation? Ask some hard questions about your concept before you test. We like this list of ten questions about strategy, especially the first few: Will your strategy beat the market? Does your strategy put you ahead of trends? TL:DR go big or stay home.  


Let’s say you’re Amazon and you want to keep growing at 20% a year. You need to find more than $50B in new revenue this year. Fifty. Billion. Dollars.

Or maybe Meta, who has made a big public commitment to the metaverse but desperately needs other new products to fill in until enough goggles arrive.

Meta’s solution? Hire entrepreneurs. Lots of them. The behemoth has been advertising for months on LinkedIn, looking for “Founders, New Product Experimentation” who will “build products that meaningfully improve people’s lives, whether by solving important social issues or delighting them as they go about their day.”

Meta has a site devoted to the NPE group which seems to give Facebook partial credit for launching ride-sharing, music streaming, and gaming industries. For its next act, however, Meta says it is hiring entrepreneurs who will put disenfranchised groups at the center of new product innovation. And they are hiring for both internal development and seeking to invest in early stage startups.

While we at Spark love the idea of experimentation, we have to ask: why in the world would you build products without knowing that there is a market for them? Indeed, “of the thirteen experimental apps launched by the NPE team since November 2019, only two still remain active,” writes Social Media Today.

To be fair, for a megalosaur with massive budgets like Meta, throwing lots of new products out there and seeing what sticks is a decent approach. It probably has some additional benefits in talent development. For more petite players, test before you build may be more cost-effective.

More Insights

September 28, 2023
Who’s paying your rent?
How smart brands find new audiences without losing their current one.
August 2, 2023
Innovate Faster—Meet Design Sprints & Heat-Testing
How design sprints and heat-testing bring speed and confidence to product innovation.
July 20, 2023
Now serving: piping hot hyper-niches
How to survive (and thrive) in a marketing world that’s increasingly niche-ified