Build Your Own House

How Much is it to Building a Building

Many hardware startups don’t realize they’re building too much. When I say “too much” I really mean two things:

1. Too many features — startups don’t know which parts of the product their target customers respond to so they develop all of them.
2. Too many products — unless founders have gone through a full product lifecycle, they usually underestimate the overhead of managing multiple permutations of a given product.

These are some of the most challenging things about hardware and few people talk about them. You must know what people love about your product before you manufacture it.

Too Many Features

Knowing which features your product “must have” is extremely difficult so many founders choose to include more features than they need. Only founders that have previously built hardware know how dangerous this is:

Each feature you add to a hardware product creates an exponential amount of work.

Founders often prioritize design and prototyping over reliability, cost, marketing, and sales during the beginning of product development. Each feature requires careful consideration because of the magnitude of additional work and complexity it creates.

Each feature of your hardware requires a huge amount of overhead. As you go up the stack, determining value becomes less precise and by most measures, more difficult. For example: determining if a feature is “necessary” seems easy but understanding how it changes the value of your brand is next to impossible (yet far more important).

If each feature takes so much effort, how do you know which features to include and which to leave out? Here are a few tips to help narrow a product down to its core features:

• Break the product down into extremely simple components. I think of product development like a high-school science project: isolate variables and be meticulous with methods. Most engineers only want to show complete prototypes to users. Break this habit. For example: if you want to test where a button should be on the handle of a product, build a few rough looks-like models of just the handle. You don’t need the entire product to test a specific problem with users.
• The magic number of people to test with seems to be around 30. Much more than 30 people takes a lot of time, but much fewer is not indicative of traction.
• Taking away features and measuring reaction is often more powerful feedback than adding them.
• Emotion trumps rationality. Find ways to measure or test users’ emotional response to a product (how they say something is more important than what they say). From the example above: users may say adding the button was fine but spend 5 extra seconds trying to figure out what the button does with a look of frustration on their faces. People don’t like to admit when they struggle to figure something out, even if their behavior tells you otherwise.

Using these tools, your feature set should actually shrink prior to a product release (rather than grow, as most products do during development):

Too Many Products

Once a product is released there is immense pressure to add additional permutations of the product (known as a stock keeping unit or SKU). Many founders assume adding an additional color or size of a product is a simple change, but just like features, adding a SKU creates overhead.

Each additional SKU requires the entire “product system” to adapt. Here are a few side-effects that can be extremely painful:

• In addition to forecasting the total number of units for the next production run, you have to figure out which SKUs are more popular and by what percentage, then forecast the number of units per SKU.
• Different distributors will sell different SKUs in different proportions based on their demographics. Kickstarter (with its predominantly male audience) will likely sell a higher percentage of black units than Bed Bath & Beyond (with its predominantly female audience), making already complex forecasting even more difficult.
• When processing returns, different SKUs must be processed and categorized separately, creating overhead for replacement units, repair parts, and return shipping.
• In addition to core SKUs, retailers often want unique SKUs for themselves. This may include colors, price points, packaging, or even features.
• SKU proliferation compounds. If you offer different sizes (small, medium, and large) of your product when it launches and then decide to add colors (red, green, and blue) your SKU count jumps from a manageable 3 to a frustrating 9.

But the biggest danger of SKU proliferation is not technical overhead, it’s brand risk:

With each product permutation, you trade accessing a larger audience for potential brand destruction.

Maintaining a focused product line is paramount to insuring your customers continue to identify and associate with your brand. If you add too many SKUs to your product line too quickly, it’s easy to confuse your customers and cause them to choose your competitor instead; Dell enabled users to have ultimate customization (an infinite number of SKUs) while Apple famously kept their SKU count low with its “2x2” methodology. Which company do you think has a stronger brand?

Balance

There are several successful consumer hardware startups that have demonstrated mastery of selecting features and extending product lines. Fitbit is one of my favorites:

Includes only unique primary SKUs (not accessory SKUs)

For its first three years selling products, Fitbit maintained an extremely focused product line of just three unique SKUs. It wasn’t until the addition of the first wrist-worn product (Flex) that the number of SKUs started to increase rapidly. By 2013, Fitbit understood supply chain management, retailer demand/forecasting, and customer support well enough to launch a product with multiple sizes and colors: but it took 4 years and selling over 1.5 million devices before they felt comfortable doing so.

Great hardware startups attain mastery of building products with only essential features, and a product line with only essential products. Finding this balance is a dangerous dance that takes continuous experimentation, meticulous planning, and an open mind.