Quality is Free


The US auto industry was turned upside down by the Japanese, who relentlessly followed Deming’s teachings. In mature industries, it’s more important that products be reliable and inexpensive than be overwhelmingly innovative.


“Quality is free.”
– W. Edwards Deming, 1968

The US auto industry was turned upside down by the Japanese, who relentlessly followed Deming’s teachings. In mature industries, it’s more important that products be reliable and inexpensive than be overwhelmingly innovative. In these industries, improving quality improves profits and market share.

“What I wouldn’t give for a hand-phaser.”
– J. Tiberius Kirk, 1968

In contrast, high tech (and what could be more tech than Star Trek) customers have valued innovation more than just about anything else. For example, if you had a phaser to sell today, the Pentagon would pay anything for it. They also wouldn’t care if it had a few manufacturing defects. The issue for this month: making the right quality/innovation tradeoffs now that the IT industry is maturing.

Quality and Stability vs. Features

In the early stages of a high-technology product’s life-cycle, the key issue is to get a usable version of the product to market quickly. Customers are willing to overlook huge quality and product-finish issues, because even a fragmentary product can give them an unbelievable profit advantage. For example, in the 90s, SGI and Sun could put out hardware with more than 1 percent of the units “dead on arrival” because the customer’s alternative would have cost millions. Intel deliberately launched their infamous “math bug” Pentium chips because they thought customers valued speed over potential inaccuracy.

In areas of hyper-innovation and disruptive technologies, the company that comes to market a few weeks earlier than competitors may get windfall profits. In these red-hot markets, competitive products bring down prices very rapidly, so product life cycles (from “widely anticipated launch” to “clearance sale”) may be less than a year.

But this cycle only works as long as the new products are better in a way that’s meaningful to the customer. Once a product category matures, time to market for an incremental improvement doesn’t matter as much. Quality and fit/finish become paramount: a product that sort-of-works just doesn’t help the customer.

When Push Comes to Shove

Introducing a new product


means making tradeoffs between features, performance, price, quality, reliability, and time to market.

Unfortunately, customers rarely give you a clear answer about these tradeoffs: you’ll never get permission to compromise on quality. But they also want their favorite new features and they want a lower price.

Until your new product comes out, each new “big deal” means additional customer feature requests. The dilemma is, you need to stabilize the product so you can move through the test and release cycle. Every new feature request makes for scope creep and another schedule slip. The delay means you will be less competitive because your current shipping product is getting older and your wonderful new product isn’t usable yet. Your weakening competitive situation increases the pressures for “just one more feature”. It’s a very slippery slope.

Controlled Chaos

Often, companies respond by creating customer one-offs that use the latest technology base, adding the particular feature needed to close the deal. At best, this approach is wasteful. Typically, however, the product is so unstable that it only convinces the customer’s techies…not the buyer. They rightly demand a stable product before they’ll pay for those special new features.

It’s almost always a better idea to create a stable base without the new features, release that as the “generally available” or “FCS” version. Customer-special features can be more easily added (and maintained) on top of that solid base.

The Train Model

It’s an even better idea to use the train model of product releases and do away with customer specials. This model is based on the idea of a train schedule: no matter how many passengers (features) make the departure time (deadline), the train will leave the station at its pre-defined time. Once customers see a predictable pattern of new releases, they begin to trust the company when it says “your new feature will make it on the next train.”

The train model can make for significant operational as well as marketing efficiencies. In the auto industry, car-model changes are traditionally made every September. From design to production to distribution to sales and even to customer purchase decisions, the annual cycle makes things easier. Most customers can’t absorb revolutionary product changes quickly, so having a regular cycle of minor enhancements (new features and fixes to old ones) helps everyone make better decisions.

In both consumer and business high-tech products, it’s really cool to have a series of incremental changes to stimulate demand. Look at what Apple is doing with the iPod — minor improvements every couple of months, with price decreases for the old models, so people get an excuse to buy now. Even if Steve Jobs had the perfect design today, it would be more profitable to introduce it over time as a series of small upgrades.

Situational Ethics

The right thing to do depends on where your market is, and what your customers’ needs and preferences are. The ultimate question is, where is your market now?
To find out, look at your competitors. Watch what customers do, as well as what they say. Talk to industry analysts and the press — is the market hungrier for the Next New Thing, or rock-solid reliability?

The bottom line: right now, most markets are tired of flaky products.

Scroll to Top