Why Do Good Products Fail?

Rajiv Gupta is a Principal Product Leader at AWS driving cross-functional business strategy for cloud infrastructure and AI storage.

getty

​When was the last time you looked at a product launch and asked yourself what that company was thinking or who in the world would buy that? I go through this at least once a week, if not more. This is in addition to the products that barely register, even after someone gives you a demo.

Many such products either die or, even worse, hang around for years without meaningful adoption. Product teams keep working on the next feature that their market research says is critical to unlocking the hockey-stick growth curve. It often doesn't deliver the intended outcomes.

A while back, I started looking for a simple set of questions that a product manager can ask when shaping a product or evaluating a direction. I reviewed both successful and unsuccessful launches across AI tools, enterprise software and consumer products. In my experience, three questions consistently revealed whether a product was set up to succeed or headed for one of the failure modes I described above. Others may have frameworks that work equally well, but these are the ones that resonate with me.

1. Can most individuals in your target segment understand the product, even if they're not the ones buying or using it?

The benchmark here isn't whether your buyer understands the product; it's whether the broader segment does. Only aerospace engineers need to understand a component built for space shuttles, not the general public. Within that segment, however, engineers should easily understand the value of the component over alternatives.

Similarly, a networking engineer should immediately understand a new switch built to handle data center-scale traffic—even those whose current infrastructure doesn't need it. They may never buy it, but they should instantly recognize what problem it solves.

The principle applies equally to consumer products, where the segment is often much broader. When ChatGPT launched, it simply said "ask anything." You didn't need to be a technologist to understand what it did, and that ease of understanding within a very broad segment was a key driver of how quickly it grew. AirPods followed a similar pattern. Any iPhone user immediately grasped what they were, even with no intention of buying them.

To be clear, this isn't about how many people understand the product. It's about whether the right people do. If most individuals within your target segment struggle to understand what a product does, that's a product problem.​

2. How much distance is there between the product's usage and value creation for the customer?

Some products make it easy to connect usage to value. When you open Claude Code and start coding, the link is direct and obvious. You don't need to run calculations, convince another team or evaluate trade-offs to know whether it's working for you.

Most products sit somewhere on a spectrum. Take a photocopier that consumes less ink. You need to assess print volumes, ink costs and projected usage before the value becomes clear.

At the far end are products with which it's genuinely difficult to assess the value. Consider a new generation of energy-efficient servers. IT buys and operates them, but the savings show up on the electricity bill that completely different people own. Additionally, how much you actually save depends heavily on your workloads. A team running sparse tasks may see negligible savings, while one running dense compute could see significant ones.

That's a lot of work before a customer can make the case for the purchase, and many teams simply won't do it.

It's worth asking where your product sits on this spectrum. The harder it is for customers to connect usage to value, the more work they have to do before they can justify the purchase. That doesn't make a product unsellable, but your go-to-market strategy and sales cycle need to account for it.

3. How much friction does the customer face before they can start using the product?

At one end, take Duolingo. You download the app, and you're learning a new language within minutes. There's almost nothing standing between you and the product's value.

At the other end sit products such as enterprise CRMs. Before a single salesperson logs a deal, IT needs to integrate systems, legal needs to approve data handling, and managers need to overcome resistance from users who are comfortable with how they currently work. Months can pass before the product is meaningfully in use.

In the middle sits something such as learning to drive a stick shift after years of driving an automatic. There's a real adjustment period, but it's manageable, and the path forward is familiar and well understood.

High-friction products require a fundamentally different go-to-market motion. CRM companies invest heavily in training programs and dedicated implementation teams. They need CXO-level alignment, top-down mandates and long negotiation cycles just to get to a signed contract. That's a proven business model, but the entire product needs to be built around it from day one.

It's time to put it all together.

These three questions aren't a post-launch checklist. They should be asked early and shape every decision, from product design to go-to-market strategy to value messaging. A misalignment on any one of them can be enough to derail an otherwise strong product.

This is also why adding more features rarely fixes a struggling product. If the segment doesn't understand what the product does, the value is too far removed from usage or the friction to adopt is too high, no amount of incremental functionality will move the needle. The product will keep hanging around without meaningful adoption, and the hockey-stick growth curve never materializes.

The products that endure tend to do well on all three. Their segment easily understands them, they deliver value close to the point of usage, and they minimize the friction between the customer and that value. I've returned to these questions repeatedly. In cloud infrastructure and AI product development, where I build at enterprise scale, getting any one of these wrong makes success very hard.​

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?