Tag Archives: startup product

No, I Can’t Give You A Roadmap For Our New Product (Yet)

Cartoon by Roger LathamA fellow product manager that’s working on a new product idea recently wrote to me:

“Common feedback I receive from our our engineers and executives is they don’t have a good grasp of the product vision. They say, “OK, that’s great, we can build that. But where are we going with this if we find the hypothesis to be true? What’s the long term vision for the product?” In essence, they’re asking what’s the end goal in 2-5 years, and if you show me that I’ll have a better sense of the architecture and tools I need to account for.”

This product manager is right at the very initial stages of their product idea, where he still needs to test the problem and solution hypotheses. But he’s already being asked for a long-term product roadmap! Sound familiar?

While the request seems perfectly reasonable, it’s misplaced at such an early stage. The question about architecture and tools seems perfectly reasonable on the surface, but it’s a scale question, and is not the right one to be focusing on before you even know if you’ve identified the right customer problem and have proof that your solution approach is viable to solving that problem. Execs are trying to assess the potential market opportunity, the underlying investment that will be needed, and the speed to achieving ROI. So naturally they want to see the long-term roadmap. Again, perfectly reasonable on the surface, but at such an early stage, you’re likely in no position to be able to answer the question.

Even at the conceptual stage, you likely have a list of potential features in your mind. You could prioritize them using one of the many scorecarding techniques written about by seasoned product practitioners. (See this, this, this, and this, to reference just a few.) These are all very valid techniques written by product folks who really know their stuff.

If you do that, though, you’ll be wasting your time. Creating a product roadmap is predicated on having a coherent product strategy, which is predicated on having a validated understanding of who are your customers, what are their pain points, and whether they’ll find your solution valuable. If you don’t even know if customers will buy your solution, what’s the point in having a roadmap?

So when do you develop a roadmap for a new product? I get this question a lot on my product help calls, so I thought I’d share my answer here.

For a startup product, the first step is always to identify the customer segment and customer problem. Quickly capture your product vision, formulate your customer, problem and solution hypotheses, and systematically test them. As you go along, you need to identify potential “innovator” customers and early adopters to whom you could deliver your solution — typically you build the product for these folks first. If practicable, test pricing at this stage as well.

Figure what you absolutely must deliver to these folks to solve their #1 problem, and work like hell to deliver it as quickly as possible. All other features get cut from scope and sit in the backlog. Again, with respect to pricing, what the customer is willing to pay to solve for takes clear precedence over anything else.

After delivering this, say, MVP, you need to actively gain feedback from these early customers. You’re using your delivered product to gain deeper insights into the customer’s problem, and you’re trying to understand what you need to improve in the product to (a) get these customers to stick, and (b) attract more new customers.

In addition, now that you have an initial set of engaged customers, you can also try to test their second level set of problems (or discover new ones). Understanding those problems may identify new enhancements/features. You’ll now be armed with a set of improvements, fixes and new ideas that you can put into the backlog.

If you have a sales force, and have armed them to sell your MVP, make sure you’re actively gathering feedback from them as well. Of course, not all their feedback may be feature-related — some may be about testing and evolving your sales messaging and positioning. However, as it relates to feature gaps, you put those into the backlog as well. I pay particular attention to feedback that’s preventing a customer sale.

You’ll have a pretty good backlog at this point, so you can now start building an initial roadmap. Start by prioritizing the backlog based on a reasonable customer-centric set of criteria that also help deliver on your company goals. I typically skew my priorities heavily toward voice of customer (VOC) feedback. While at any stage of the product lifecycle, features should solve tangible customer problems, it’s even more important at this early stage.

I also factor in the company’s strategic goals. For example, if the company’s focus is retention, features that create stickiness may carry more weight; if the focus is growth through customer acquisition, then sellable features may be more important; if it’s revenue through customer penetration, features that drive engagement and up-sells may take priority.

I also make some allowance for operational issues. I may not necessarily have a scale problem (yet), so these type of issues may not take precedence over VOC or driving revenue; however, I don’t want to completely ignore technical debt or reasonable operational fixes.

Once you have a prioritized list, socialize it. (Read this post by Bruce McCarthy on using “shuttle diplomacy” to get buy-in.) For the top priority items on the list, get t-shirt sizing from Engineering, and make a final call to sequence out the items based on customer value vs. LOE. Now you’ve got a validated product in the marketplace with a decent first-pass roadmap that you can build upon.

5 Steps To Validate Your Product Idea Without A Product

Here’s a scenario:

Top Exec at your company comes to you and says, “Yesterday, I was talking to Big Industry Player and they mentioned how they have Shiny Object. I think we should have Shiny Object too. How fast can we get it done?”

Sound familiar?

At last weekend’s ProductCamp DC event, my co-founder and I hosted a session called “Tales From The Product Frontlines”. Our first topic was about this very scenario, and it so energized the participants that it took up 30 of the 45 minutes we had been allotted!

During the session, we asked product people questions like:

  • What do you do now to vet new ideas, whether your own or from some other source?
  • How is that working?

Most folks talked about having some sort of governance structure, such as an Executive Steering Committee to vet new ideas based on some established criteria. Many talked about the need to create a business case, and some advocated that Product Management is best suited to do that, regardless of where the idea came from. Yet, after the session, every person I spoke with told me they felt writing a business case was a total waste of time.

Why is this? Because the old ways just don’t work.

Pragmatic Marketing’s 2013 survey revealed that product managers spend over a month’s worth of time writing business cases. These business cases are filled with highly assumptive 3- or 5-year projections that are used to support significant investment asks. It’s no wonder we hate this — we’re staking our professional credibility on these unvalidated assumptions!

The problem is we were never taught a systematic method by which to obtain the crucial information needed to inform a business case.

So what’s the solution?

Let me pause here to say if you’re hoping I have a magic secret for quickly validating a product idea, or if you’re totally married to writing business cases or MRDs, then stop. This post isn’t for you, because:

  • It’s different and requires you to THINK.
  • It’s hard work.
  • It can take some time to get results.

But it works. I’m writing for the 10% who (1) realize I’m telling the truth (that the old ways just don’t work), and (2) are willing to try something different.

(Thanks, Kevin Dewalt, for putting this so well, and forgiving me a little plagiarism!)

The solution is validated learning

Bottom line is spending time writing a big document is a colossal waste of time. Instead, focus on validated learning.

Because any new product idea is based on assumptions, those assumptions need to be validated. This can be achieved by formulating testable falsifiable hypotheses around the riskiest assumptions, and rigorously testing these hypotheses.

The process of validation can be outlined in 5 steps. Each step generally follows this meta-pattern:

  • Formulate a testable falsifiable hypothesis.
  • Test the hypothesis.
  • Analyze results and learnings.
  • Decide to pivot or persevere.
  • Repeat.

I’ve used these this process to validate ideas for software products, but I imagine it could be adapted for any product concept. (So give it a try and let me know your results!)

1. Write down your customer hypothesis.

Most folks typically start with the solution first, which is the wrong place to start.

You need to take a step back and think very deliberately about your customer. This is true whether you’re pursuing adding a new component to an existing product, or are pursuing a truly new product.

Note, by customer I mean the individual who will buy your product. Even in B2B, you need to think about the specific individual or set of individuals to whom you will need to sell your solution. That’s your customer.

2. Write down your problem hypothesis.

What problems does your idea solve for the customer? One way to do this is to think about the goal or job the customer is trying to accomplish. For example: “I believe [customer] has a problem achieving [goal].” Or: “I believe [customer] is trying to accomplish [job], because [desired benefit].”

I typically use the Product CanvasTM to do this, as it allows me in a focused manner to break down the problem domain into discrete problems, and then formulate hypotheses around what I believe to be the top problems that my solution absolutely has to solve for first.

3. Validate your customer/problem hypothesis.

Now it’s time to test your hypothesis. You do this by talking to folks you believe meet your target customer demographic.

Be sure to define a minimum success criteria, which is the minimum amount of data you will need from the test to justify investing more time, effort or resources into proceeding with the idea.

When you’ve met your minimum success criteria, analyze your data, and decide whether to pivot or persevere.

A pivot is a fundamental change in direction of your business model or product strategy. You face a pivot when your hypothesis has been invalidated (i.e., proven false).

At this early stage, you could expect a pivot in terms of a change to your target problems, your target customer, or both. This may mean going back to step 1 or 2.

However, if the results of your test prove (i.e., validate) your customer/problem hypothesis, you may decide to persevere, and move on to step 4.

4. Validate problem/solution fit.

Now that you’ve validated your customer/problem fit, you need to test whether your idea is a potentially viable solution to the customer’s problem.

There are many ways to go about this, but nothing really beats creating a visual representation of your envisioned solution in the form of a wireframe or mockup. To keep things simple and minimize work, I look to design a representational screen or flow for each discrete problem identified on my Product Canvas and validated via the previous steps.

The primary goal of this stage is to garner early customers who endorse your solution vision, and would be willing to use (and ideally pay for) an early version of the product. You’re also looking for directional feedback to identify the “right” handful of features to build for these early customers to prioritize your product development.

Formulate hypotheses around your screens, define your minimum success criteria, then reach back out to the customers you had interviewed earlier and demo the screens to them. Also try to mix in some new folks who fit your target customer demographic.

When done, just like you did at the end of step 3, analyze your data and decide whether to Pivot or Persevere. At this stage, you may pivot on the solution, the problem or even the customer. Or, if you’ve validated your hypothesis, you may have problem/solution fit, and can move on to step 5.

5. Validate your solution via an MVP.

There are many misconceptions about what constitutes an MVP, and I’ve written about these before. In short, an MVP is an actual product that attempts to deliver real value to customers.

It’s “minimum” in the sense that it’s an attempt to deliver the absolute necessary set of features or capabilities needed to solve the customer’s problem for which the customer will pay.

A primary outcome of the previous step is being able to understand your customers’ problems at a granular level, which helps prioritize the initial set of features to build. This drives the definition of your MVP.

Once again, formulate a set of testable hypotheses to ensure you’re continuing to drive your product development based on validated learning. Depending on the complexities of the problem and solution, and the nature of my target customer, I may opt to first demo the MVP before actually delivering it. The results of your MVP test will again determine whether to persevere or pivot.

If you’ve got this far, you’ve validated critical components of your product idea. It doesn’t mean your idea is guaranteed to succeed, but you will have gained a far better understanding of the market opportunity, allowing for a far less assumptive and much more robust business case for your new product idea.

Using Customer Development To Create The Business Case For Your Product Idea

Many of my product help calls are from folks frustrated with being able to pursue a new product in an existing company.

They complain about how difficult it is to secure resources and garner internal stakeholder buy-in.

If you find yourself in this position, congratulations!

As painful and frustrating as it is, many successful product people I’ve met have gone through exactly what you’re experiencing — including me!

The problem is the way we’ve been taught to pursue this — to first write a business case, business plan or MRD — just doesn’t work. The fact is, people think writing a business case is a waste of time and hate it.

And no where are we taught how to cultivate stakeholder support.

Like it or not, every project in an existing company, regardless of the size of the company, needs an internal champion or sponsor.

Lack of stakeholder buy-in can be the biggest impediment to your product no matter how good your product idea may be.

So it’s no surprise really that product people are frustrated.


So here’s a new process I’ve been following that’s worked much better for me:

Using Customer Development To Create The Business Case For My Product Idea

In a nutshell, I deferred asking for major dollars and resources until I absolutely needed to. I used a series of validated learning milestones to build momentum internally and to build the case for investing in my product idea.

Here’s what I did:

1. I quickly sketched out my product strategy on a 1-page Product Canvas. No wasting time writing a multi-page document no one is going to read.

2. I decomposed the product strategy into critical learning milestones meant to answer the most important questions in my product strategy:

  • How does our target customer describe the problem?
  • How are they solving it today?
  • Why is that solution not working for them? In other words, why is the problem still a pain for them?
  • How can we know before we invest a lot in development, sales and marketing that the solution we’re thinking of building really solves the problem?
  • How quickly can we get our first customer?
  • What are the most important features we need to have in our go-to-market product?

3. I figured out what’s the least amount of work I need to do to maximize my learning for each milestone.

4. I broke down my investment need into these milestones, showing how ROI could be tangibly achieved based on measurable results.

Here’s what the investment plan looked like:


In the past, I would have asked to spend money on 3rd party market research (four to five figures), a design agency to craft the user experience (five to six figures — ugh!), a usability study (five figures), and a large development team (many figures).

This would be costly and take a long time before I would have delivered the product to a single customer.

And it’s a tough business case to make.

Instead, because I had broken down the plan into these learning milestones, I was able to easily accomplish the first two milestones by spending little to no money at all.

The first was simply my time, so required no money.

To validate our solution hypothesis, I used Balsamiq to sketch a handful of the key screens of our solution. Total cost was $79 for the tool and my time.

When we were ready to design the user experience, since we didn’t have an in-house designer, we commissioned a cracker-jack freelance designer — way cheaper than hiring an expensive design agency, and way faster. It got the job done.

(If you happen to have an in-house designer or design team, awesome — use them. Your investment “ask” may only be some of their time.)

In this way, I was able to use the resources I had at my disposal for as long as I could to create traction.

This approach not only allowed me to conserve precious funds and resources, but also allowed me to be less assumptive and more data-driven in identifying my investment ask at subsequent stages.

It also enabled me to not only build early traction with customers, but also have them help me define the minimum feature set we’d need to develop to go to market — labeled as the Minimum Sellable Product (MSP) in the picture above.

Here are the benefits that resulted:

  • Instead of writing a massive business case based largely on guesswork, I needed only to sell a bunch of mini-business cases. Way quicker and easier to do.
  • Each mini-business case was informed by the learnings from the previous stage, making each subsequent mini-business case better informed, more robust, and an easier sell.
  • The product strategy was informed by real market insights. (What a product manager needs to do anyway!)
  • I had a customer driven product roadmap that was tough for anyone to dispute as it was informed directly by tangible customer insights, which defined what went in our MVP vs. MSP vs. roadmap vs. nice-to-have.
  • This enabled our product development efforts to be more focused, as I had all the ammunition I needed to fend off arbitrary new feature requests that risked derailing our product development.
  • Because of our “co-innovation” approach with our customers, we were able to get “earlyvangelists” that we could leverage to generate momentum for our broader market launch. Customer Development in concert with Product Development!
  • All of this made it much easier for me to garner, maintain and accelerate buy-in from my internal stakeholders, because:
    1. My plan showed a clear milestone-based investment plan with the ROI to be gained at each phase.
    2. Smaller continual investments are easier to digest and support than a large upfront one.
    3. Each investment stage was grounded in real customer data, increasing confidence in pursuing the product.
    4. My stakeholders felt involved in the process, as I made sure to keep them informed and provide them an opportunity to provide feedback.
    5. This, in turn, kept me one step ahead of any potential concerns they may have had, and I could make sure to address them at a future stage.

An MVP Is Not The Smallest Collection Of Features You Can Deliver

Source: Spotify

Source: Spotify

There’s a lot of discussion and confusion about what is and isn’t a minimum viable product (MVP).

Worse, many execs have latched on to the term without really understanding what truly constitutes an MVP — many use it as a buzzword, and as a synonym to mean a completed version 1.0 ready to be sold to all customers.

Buzzwords are meaningless. They represent lazy thinking. And using “MVP” to mean “first market launch” or “first customer ship” means you’re back to the old waterfall, traditional project-driven software development, sales-focused approach. If that’s your approach, fine. Just don’t call what you’re delivering an MVP.

On the flip side, lots of folks in the enterprise world, including in product management, over-think the term. It gets lost in the clever nuances of market maturity, and a long entrenchment in the world of release dates and feature-based requirements thinking.

Many folks think of MVP as simply the smallest collection of features to deliver to customers. Wrong. It’s not.

The problem with that approach is it assumes we know ahead of time exactly what will satisfy customers. Even if we’ve served them for years, odds are when it comes to a new product or feature, we don’t.

Now, the challenge with the concept of a minimum viable product is it constitutes an entirely different way of thinking about our approach to product development.

It’s not about product delivery actually — in other words, it’s not about delivering product for the sake of delivering it or to hit a deadline.

An MVP is about validated learning.

As such, it puts customers’ problems squarely at the center, not our solution.

Reality check: Customers don’t care about your solution. They care about their problems. Your solution, while interesting, is irrelevant.

So if we’re going to use the term “MVP”, it’s important to understand what it really means.

Fortunately, all it takes to do that is to go back to the definition.

Download The Handy Primer “What Is An MVP?” >>

Minimum Viable Product (MVP) is a term coined by Eric Ries as part of his Lean Startup methodology, which lays out a framework for pursuing a startup in particular, and product innovation more generally. This means we need to understand the methodology of Lean Startup to have the right context for using terms like “MVP”. (Just like we shouldn’t use “product backlog” from Agile as a synonym for “dumping ground for all possible feature ideas”.)

Eric lays out a definition for what is an MVP:

“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.”

Eric goes on to explain exactly what he means (emphasis mine):

MVP, despite the name, is not about creating minimal products… In fact, MVP is quite annoying, because it imposes extra overhead. We have to manage to learn something from our first product iteration. In a lot of cases, this requires a lot of energy invested in talking to customers or metrics and analytics.

Second, the definition’s use of the words maximum and minimum means an MVP is decidedly not formulaic. It requires judgment to figure out, for any given context, what MVP makes sense.

Let’s break this down.

1. An MVP is a product. This means it must be something delivered to customers that they can use.

There’s a lot that’s been written about creating landing pages, mockups, prototypes, doing smoke tests, etc., and considering them as forms of MVPs. While these are undoubtedly worthwhile, and certainly “lean”, efforts to gain valuable learnings, they are not products. Read Ramli John‘s excellent post on “A Landing Page Is NOT A Minimum Viable Product“.

A product must attempt to deliver real value to customers. So a minimum viable product is an attempt — an experiment — to deliver real value to customer.

Which leads us to…

2. An MVP is viable. This means it must try to tangibly solve real world and urgent problems faced by your target customers. An MVP must attempt to deliver value.

So it’s not about figuring out the smallest collection of features. It’s about making sure we’ve understood our customers’ top problems, and figuring out how to deliver a solution to those problems in a way that early customers are willing to “pay” for. (“Pay” in quotes as it depends on your business model.)

If we can’t viably solve early customers’ primary problems, everything else is moot. That is why an MVP is about validated learning.

3. An MVP is the minimum version of your product vision. A few years ago, I had to build an online form builder app that would allow customers to create online payment forms without the need to write any HTML or worry about connecting to a payment gateway. Before having our developers write a single line of code to build the product, we first offered customers the capability as a service: we would get their specs, and then manually build and deliver each online payment form one-by-one, customer-by-customer. Customers would pay us for this service.

This “concierge” type service was our MVP version of our product vision. Of course, it wasn’t scalable. But we learned a heck of a lot: most common types of payment forms they wanted, what was most important to them in a form, frequency of wanting to make changes, reporting needs, and how they perceived the value of the service.

We parlayed these learnings into developing the software app itself — which, by the way, we delivered as an MVP to early customers to whom we had pre-sold the software product. (Yes, we delivered two different types of MVPs!)

Whether you take a “concierge” approach or your MVP is actual code, it most definitely does NOT mean it’s a half-baked or buggy product. (Remember viable from above?)

It DOES mean critically thinking through the absolute necessary features your product will need day 1 to solve your early customers’ top problems, focusing on delivering those first, and putting everything else on the backlog for the time being. It also means being very deliberate about finding those “earlyvangelists” that Steve Blank always talks about.

Ultimately, the key here is “maximum amount of validated learning”. This means being systematic about identifying your riskiest assumptions, formulating testable falsifiable hypotheses around these, and using an MVP — a minimum viable product version of your product vision — to prove or disprove your hypotheses.

Now, validated learning can certainly be accomplished via a landing page, mockup, wireframes, etc. And it may make sense to do these things. Super. But don’t call them MVPs, because while they may deliver value to you and your product idea, they’re not delivering actual value to the customer.

At the same time, the traditional product management exercise of identifying all the features of a product, force ranking them, and then drawing a line through the list to identify the smallest collection to be delivered by a given timeframe is not an MVP. Why? Because this approach is not predicated on maximizing validated learning. If you’re going to pursue this approach, go ahead and call it Release 1.0, Version 1.0, “Beta”, whatever. But don’t call it an MVP.

An MVP is about not just the solution we’re delivering, but also the approach. The key is maximizing validated learning.

I’ve created a handy primer on what is a minimum viable product. Download it below. I hope it helps you to become a pro at defining an MVP for your next great product idea!

Download The Handy Primer “What Is An MVP?” >>