The Myth of the MVP
Minimum viable products are a myth.
MVPs sacrifice craft and refinement in a rush to learn from an incomplete product that is not remotely ready to solve a customer’s problem. It’s clear how we arrived here. Anxious, harried startup founders reacted to the impatience of their investors, stopped short of “done,” and rushed to collect feedback to understand whether a product should be built. The approach robs them of the opportunity to learn the answer.
MVPs worked for Eric Reis, who wrote a book, which we rushed to emulate, tumbling over ourselves to tune an engine we hadn’t yet built.
“First, a definition: the minimum viable product is that version of a new product which allows a team to collect the maximum amount of validated learning about customers with the least effort.”
Cool.
Do iterative and incremental development. Iterative because you build features in parallel. Incremental because you complete the work in the smallest possible batches. Show it to customers when it’s ready for their feedback.
Powerful concepts. Build small, in parallel, and learn early and often.
Form a hypothesis and refine your product.
We’ve lost track of what it means to refine and resort to shipping sooner than we should. Even Reis admits people misunderstand minimum viable products. Still, the industry ran with it, and during the chase, we lost track of what it meant to refine.
The Lean Startup methodology surmises finished products are built without talking to customers. It’s an odd assertion. A few poorly conceived startups fell prey to hubris and shipped without knowing their customers. How common is this tragedy? Common enough to rearrange an entire industry’s approach to building software?
No good product manager disappears into a myopic haze to “spend months, sometimes years, perfecting [a] product” without validating ideas, gathering feedback, and knowing the problem to be solved. Most are obsessed with how their product solves their customer’s problems. Showing the customer a product not remotely fit for purpose–here’s our skateboard–won’t answer whether you’ve solved the problem, nor will it relieve Reis of his “extreme uncertainty”.
Extreme MVPs
Pockets of the industry practice what I term extreme MVPs. This meme sums up the mentality.
“Don’t build a car one step at a time,” they say, “even when you know that’s what your customer needs. Start with something super small, like a skateboard, learn from that, throw it away to build a scooter, learn from that, throw it away, and eventually you’ll deliver the car your customer wants to buy.”
Your customer told you she has a problem solved by the car. Believe her.
If she wanted a skateboard, she’d ask for one.
(Forget Henry Ford’s quote about customers asking for faster horses. He said breeding faster horses wouldn’t lead to innovations like the automobile, expressing an entirely different concept. Ford’s customers were paying to get locomotive engines off fixed tracks.)
Your users are training to be drivers. Your partners are building roads. Your competitors are building cars.
What do you expect to learn from a minimum viable skateboard?
More importantly, why do you believe your team is only capable of building a skateboard?
The remedy
The wisdom conveyed by Lean Startup tells us to put the least possible effort into our craft, resulting in raw, unfinished MVPs. Only these MVPs teach us whether we’ve solved our hypothesized problem.
I offer an alternative.
Last night, I finished shaping my first surfboard. Finished, in this case, means I carved a raw foam blank until it looked ready for glassing–covering the surfboard’s foam core with fiberglass cloth and tinted resin.
The trick is knowing what ‘ready for glassing’ looks, feels, and sounds like. Good shapers know when to stop sanding. Near the end, the increments are fine, a swoop here, a swish there. Then you stand back and scan for shadows and subtle dips, watch for curves, check each rail’s roundness, and feel the board’s thickness is consistent. Take another single pass with your sanding block. Make the board ready for subsequent refinement.
This is not an MVP. This board isn’t even prototype-ready yet. The raw foam is vulnerable to nicks and gouges easily. It certainly isn’t ready for the raw power of the ocean.
Iteration One
After each run with the planer, sanding blocks, screens, and other shaping tools, we’d stand back, observe the mess, and crack the same joke: “Glass it!”
The board looked like this at the end of the first night’s shaping.
Shipping in this state is ludicrous. And that’s what made it funny to shout “glass it” over the noise of the air compressor. The board obviously won’t work. And yet, we made incremental refinement to a surfboard you’d be comfortable riding in pumping 4-6 foot swells at your local break.
We often confuse this state for our software’s MVP. Ship it!
Iteration Two
Another iteration of shaving, planing, and sanding.
Now, you could slap on some fiberglass and resin, route some fin boxes, and attempt surfing. As an MVP, this refinement is even worse because–ignoring the square rails, nose, and tail–it looks like a surfboard, so you might expect it to work when you paddle for a wave. You’d not learn any meaningful lessons from this iteration. Certainly not what to do next to refine the shape or contours of the board.
Iteration Three
Shaping is nearly complete.
But only almost. The rail bevels are taking shape, a critical step in making the rails round and keeping the board’s curve. The bottom contours are in place, and the nose and tail are thinned out. The deck needs more sanding. On close inspection, the rails have “wobbles,” the thickness is mismatched on either side of the nose (see the shadow?), and the wooden stringer needs further shaving.
The board looks like it might work, so the disappointment in this MVP would be even greater.
Iteration Last
Finished at last.
This is our best-case MVP. Is it surfable? Nope.
Shaping is a craft. And that part of crafting this surfboard is done. We made constant refinements until no more were possible.
But don’t confuse your shaped blank for a shippable product.
Now, other craftspeople will take their turn glassing, tinting, and sanding. Then, the board surpasses any MVP because it’s usable. It can be tested against its purpose. The surfer determines fit and function. The shaped foam blank cannot answer questions about how the board paddles, turns on the face of a wave or speeds down the line.
To refine means to improve through small changes. Extreme MVPs don’t refine a skateboard into a car. There’s no series of small changes when the feedback is “it needs an engine.” MVPs force the builder to make leaps, discard what’s been built, and start over.
MVPs are supposed to tell us whether the product should be built. If you never proceed past “minimum viable,” how will you know?
Make better software
Something’s off when we collectively agree that practicing our craft and emphasizing quality matters less than expending less effort and going fast to create something minimally viable when our judgment tells us it doesn’t serve our customer’s needs.
Spend the time to deeply understand the customer’s problem and refine your product until you’ve solved that problem. Don’t repeatedly build and discard. This wastes your customer’s time, burns cash, and ruins the pride we keep for our craft.
Refine your product until the time-to-value is ten minutes or less. Feel proud when your customer experiences value quickly, knowing you spent the time to get it right, even if that meant skipping the mythical MVP.
The other MVP issue is that a lot of MVP’s become the product - execs think “Hey, it’s working well ENOUGH, let’s move resources to the next shiny thing, and we will come back to this later,” and later never happens!