The Minimum Viable Product (MVP) is a key lean startup concept popularized by Eric Ries. The basic idea is to maximize validated learning for the least amount of effort. After all, why waste effort building out a product without first testing if it’s worth it?
So for a new product idea or concept, what is the absolute minimum viable product, and how do you go about building it?
For Timothy Ferris, his MVP for testing new products that don’t yet exist (micro-testing) comprises a landing page, signup page, and Google Adwords to drive traffic. However, this approach presupposes that:
- You can create a good landing page
- You can write good AdWords copy
- Adwords is a viable distribution channel for your product
Unlike a book title or some other physical product, startups are usually characterized by products where the problem and solution are unknown and have not yet been validated, which makes writing good landing page copy hard, and good Adwords copy even harder (you only get 25 characters for your headline!). At best, you can guess. But starting with that approach is a surefire way of dumping a lot of money on Google Ads fast. Plus, the return on learning is low — When your click-through rate is low or the bounce rate is high, you get zero visibility into why. Was it poor copy, poor product/market fit, or both? And don’t even get me started on how expensive CPCs have gotten in competitive markets.
For me, the minimum viable product goes back even further to Steve Blank’s concept of the “Customer Problem Presentation” outlined in his book “The Four Steps to the Epiphany.” The Customer Problem Presentation is a scripted interview with your target customers done either face-to-face or over the phone. During the presentation, you first outline the top 3 problems you are addressing, the current solutions to those problems, and then your solutions to those problems. You pause after each section to give the interviewee lots of room to talk, as you’ll be listening more than talking here. You are listening to understand the customer’s worldview. You are listening for specific keywords and vocabulary. Lastly, you are prioritizing the problem/solution under “must-have,” “nice-to-have,” and “don’t care.” This is also a time to test pricing. You are charging for your product, right?
It is very important to use the customer problem presentation as an exploration of problems and not solutions. This is NOT a sales pitch. You don’t know what to pitch yet. It is best to let customers freestyle and go on rants. Rants are a great indicator of a pain point worth addressing. A lot of people like using surveys instead, but I don’t. They are definitely easier to create, but surveys are too rigid and assume you know the right questions to ask. But more importantly, they don’t let you listen to the customer.
For CloudFire, figuring out our initial target customers was easy. Technically, it was a pivot from a previous product, BoxCloud, which started as a general purpose file sharing service but got narrowed down to file sharing for small businesses and freelancers. I had a personal itch to test a p2web solution in the consumer space, and having kids presented just that opportunity. My wife, Sasha, who does most of the photo/video sharing, was having difficulty finding a new photo and video sharing service that met all her needs, which had changed after kids. That prompted me to ask a bunch of questions. The basic problem we uncovered was that all of the existing sites required you to babysit the sharing process, which can be a hassle. The problem new parents face is not a shortage of options but a shortage of time.
I decided to follow this trail and see if this conclusion was limited just to Sasha. I filled out a set of hypothesis worksheets in Steve Blank’s book on the product, customer, channel pricing, demand creation, market type, and competition. I would recommend everyone formalize this process. My initial scan of the worksheets made me believe I already knew all the answers. I involved Sasha in the process, and discussions that I thought would be 30-minute conversations turned into 2-hour discussions as she questioned almost all my assumptions… Yes, I still love her after that… Following a customer development process, the biggest mind shift is from thinking you know something to testing everything you know.
We built out our initial customer problem presentation and decided to target people just like us — busy parents with young kids.
Our top 3 problems were:
- Sharing lots of photos and videos is a hassle
- A lot of services downsize the images, so the quality is poor
- Notifying family and friends of updates was manual and a chore
We were able to find the initial batch through friends and daycare and subsequent batches through follow-on referrals. I’ll add that it is very important to talk to strangers to keep objectivity in check. Family and friends can be too kind sometimes and lead you down the wrong path. We debated paying for their time with gift cards or doing a DSLR camera raffle and, in the end, decided just to lay out our objectives and ask for 30 mins of their time. That was enough. The hardest part was scheduling time to talk. The fact that we were going after busy parents didn’t help. We kept track of our target list, response rates, presentation scripts, and customer responses on our internal wiki.
During the interview, we were particularly interested in learning what their sharing workflow was like. We set up the stage and let them tell us everything they did with their photos/videos, taking them from camera to shared, what they wished they could change, and the magical pricing questions: Would they use a solution like the one we were envisioning if it were free? Would they use it if it were $X/yr? X changed from customer to customer, but we kept it as real as possible.
We talked to enough people until their answers started sounding the same. At that point, we had a pretty good idea of what our product’s unique value proposition should be, a list of other benefits, and a price to put on our signup page.
Our revised top 3 problems were:
- Sharing lots of photos and videos is a hassle (stayed the same)
- Requiring visitors to signup is annoying
- Photo Gallery design was too busy or complicated
From here, we had two options: We could either build landing and signup pages and start testing without a product. Or, we could start building a minimal product. There has been much debate on the pros and cons of each approach.
While leading users to a signup page with no product behind it can help test the messaging, it can also hurt your trust and credibility. If you charge for your product, this is even shadier and on the verge of being illegal. If the traffic volume is low (which it will be at the beginning), you can get away with this. But I fear that too many people are starting to test this way. If enough users experience these fake signup pages, there will be a backlash of some sort.
We were lucky that CloudFire was an offshoot product, and I could very quickly (in a couple of weeks) create a usable product that I could offer. As soon as that was ready, we hit the same potential customers again for a product presentation. This time around, we presented the revised top 3 problems we were addressing and then proceeded to show them how we solved them with our solution. We asked them how they would describe what was different about our service, which gave us valuable keyword and positioning statements. The call to action at the end of the presentation was for them to signup for the service. Another great side effect of this process was that I ended up with a scripted demo that I easily turned into a screencast and eventually used on the landing page.
Face-to-face selling is definitely not scalable for a product like this. Still, engaging customers right from the beginning helped us to strengthen our relationship along the way, which made them even more willing to help later.
I then *finally* turned my attention to the landing and signup pages. It did take a lot more time to get here, but my gut feeling from the start was that building good converting landing pages was going to be a challenge, especially in such a competitive existing market. I also had some serious questions about whether Adwords would be the right distribution channel. Everyone we interviewed had found their existing solution through a referral from someone else and not through Google. I decided to test Adwords, StumbleUpon, and Facebook anyway. More on how I built and tested my landing page(s) next time.