(this post originally appeared on the Neo blog)

Andy Weissman of Union Square Ventures wrote a piece the other day on chaos theory and startups. His conclusion was that making decisions, and deciding how you make decisions, is of the utmost importance.

There are many approaches to making decisions. My current thinking can roughly be described as “lead with vision, but reality check with experiments.” I use frameworks such as jobs-to-be-done, design thinking and lean startup to help ground me. But that is not the only way.

Regardless of method, the need to make smart, fast decisions is universal.

This fact applies to new products as well as new companies.

I was speaking to an executive at a well-known company about team structure for an innovation group. The question was whether you could have one team validate an idea, and then shift the work over to a “production” team to implement and ship.

I argued that you should have at least *some* of the initial team remain with the product, especially the product manager/leader. When you emerge from an initial vetting phase for a product, the decisions are just beginning, not ending. If you have not spoken to customers directly, if you have not seen the results of your experiments, and if you have not experienced the nuances of your value tests, how can you make good decisions? How would you know what was important, what to cut or compromise, what to keep, and where the big risks remain that need further learning? Splitting the teams sounds like a sure-fire way to get bloat in your production work.

When I say bloat, I am talking about costs as well as features. This is true even if the people on your production team are cheaper. I’d rather have a small, elite team punch out something focused and valuable, than hand a fledgling product over to a confused group. Confusion — the inability to make fast, well-informed decisions — leads to wasteful time delays, feature expansion, maintenance headaches and ultimately a customer who is as confused as the team itself. Or more likely, a non-customer.

Lastly, an important part of Andy’s message is about speed, not just about the quality of decisions. You need both. So the key is to find decisive women and men who are comfortable balancing intuition, data and the wisdom of their team. Then you need to keep them involved in the entire lifecycle of getting something off the ground.

Powerful new ideas often start with a spark insight: “Carpooling sucks! How can we fix it?” Or “Wow, these new smartphones might allow me to disrupt the entire taxi industry in a way never before possible!”

Lean is great, but before you jump ahead with customer development, experiments and MVPs, it is worth taking two preliminary steps: 1. examine the competititive landscape; 2. ask the question, “does this feel big enough to work on?”

I want to focus on the latter here. It’s not about defining a market size in Excel. That can be straightforward if you entering an established market, but very hard if you are trying to create a new market (see Blank). But it is worth taking an educated guess on the addressable market, and examining if you, as the entrepreneur, can connect the dots from today to a huge opportunity.

Part of this process, which feeds directly into customer development work, is to segment the market. You want to get as specific as possible and work your way down.

For example, “We want to create a SaaS time tracking application for small and medium businesses” (think Harvest). Is there a real business there?

Small and medium businesses is too vague. Let’s go one level deeper:

1. SMBs that pay their people based on hours worked (restaurants, building contractors, florists, etc)
2. SMBs that charge their customers based on hours worked (law firms, building contractors, childcare agencies, etc)

Already we can ask interesting questions — can one product eventually serve both segments or are we talking separate products or even businesses? Is one of those a better first market for us? Can we see a path to jump from one segment to the next? The answers will just be guesses, but take your best guess.

Next you want to look at your segment and do the same thing with the subsegments. Segment #2 has two big buckets:

2.1 knowledge worker services such as law, engineering, creative/design, accounting, strategy, and others.
2.2 more “blue collar” services such as construction, handyman services, childcare agencies, and others

Do they have enough in common that we can view them as one segment? In this case, we feel that the business models, sales cycles, etc might differ dramatically between the two, and we should focus on knowledge workers.

Again, ask if we can view our “knowledge worker” segment as one group that can be reasonably addressed by one company, or does it need to be segmented again?

Let’s make the assumption that we can group knowledge workers together as a single addressable market. Oh sure, there will be some differences — for example you’ll have a different sales cycle and more entrenched competitors in law firms — but for the purposes of a swag, we can group them together.

If we look at NAICS codes 54 (Professional, Scientific, and Technical Services) we see that there are over 1.1 million establishments in the USA and over 8.5 million employees.  We don’t know how many of those are time-and-materials vs fixed cost business models, but let’s swag and say 50% (550K).  If you can get roughly $70 a month on average from each firm (obviously an assumption to test), you are getting close to a $500M-a-year total opportunity for the US market (if you had 100% of the available market, which of course would not happen).

This might tell me that it has the potential to be a decent but not earth-shattering business. As conceived, it is probably not one that a typical VC would invest in, but it could be a good and sustainable business. Whether that is good or bad depends on the my (as the entrepreneur) context and desires. And maybe we would come up with ideas that could crack open new value to take it to an entirely new level. Google certainly did.

By taking a look at realistic segments, you can take a slightly more informed swing at market size which would then inform questions like: is this worth my time? what kind of financing could I, or should I, get for this kind of idea? how should I balance growth versus sustainability?

It’s also a very important thing to be able to answer when you are fundraising, although as Bryce writes, it’s probably more important that you have a thoughtful answer than whether you can convince anyone with data.

With this work behind you, you can now take your segments and start digging into customer needs and their associated risks, and get off and running with customer development, more specific personas (especially around early adopters), lean experiments, and creative product thinking.

Custdev – Everyone Does It. Full Stop.

by Giff on February 10, 2015

I see a lot of product teams try out customer development at the beginning of their product journey. But quickly the team gets so busy shipping features and they outsource customer learning to customer support, research or usability specialists, or the sales force. Those are all valuable touch points, but custdev is not a task that product teams should ignore.

Why do we do qualitative research?

  • it lets us test value (very different from usability) and whether a problem is really a problem
  • it builds deeper empathy with the customer
  • it creates shared understanding across the team
  • it tells us the “why” behind customer behavior
  • it sparks our big insights and mental leaps

Why would you ever want to stop doing any of that?

My colleague Jeff Gothelf correctly demands that *everyone* take part in customer research. Even the introverts who hate it — let them be the note-taker. Even the boss who wants to delegate everything — they only have to do a few sessions.

People only believe what they see with their own two eyes, ergo they need to directly participate in customer development. Direct learning means more belief in the learning. More belief means fewer arguments fueled by ego rather than data. And ultimately, shared understanding equals more speed, better teamwork, and more informed decision making.

Jeff, Josh Seiden and I spent last week with the teams at AutoTrader in the UK. They are already a very sophisticated product group, with cross-functional teams and advanced dev-ops. They were already talking to customers, but not usually as an entire team. I can’t tell you how many people have come up to me glowing with the benefits of exposing everyone to the customer. They are feeling faster, more creative, more grounded in real needs and ultimately more effective.

Don’t outsource your custdev work, and make sure everyone takes part.

If you are interested in learning more about customer development, check out Talking to Humans

The Value Behind Faster Horses

February 5, 2015

David Bland has a funny, and painful, graphic called the Product Death Cycle. It clearly illustrates one of the more dangerous parts of customer development: taking customer ideas literally. The famous line attributed to Henry Ford goes, “If I had asked my customers what they wanted, they would have asked for a faster horse.” Some […]

Read the full article →

Lean Startup Doesn’t Need Tools

January 29, 2015

In 2012, we were testing a bunch of new business ideas for Amex and I thought, “what if we had a tool that let us capture our assumptions, our experiments, and our progress?” So over a couple weeks, we hacked something out and tried it out on a few projects. It was a complete waste […]

Read the full article →

Death to Clickbait

January 28, 2015

I’ve decided I’ve had it with blogger clickbait, in particular the tendency to toss a highly controversial title in front of a non-controversial post. The ones I fall for usually claim that something I believe is wrong (or “dead”). Then they proceed to explain how the idea was right after all, albeit in need of […]

Read the full article →

Making Innovation Unsexy

January 21, 2015

(This post was originally posted on the Neo blog) “None of my inventions came by accident. I see a worthwhile need to be met and I make trial after trial until it comes. What it boils down to is one per cent inspiration and ninety-nine per cent perspiration.” – Thomas Edison, 1929 Many years ago, […]

Read the full article →

Startups: Don’t Wait to Build Exit Relationships

January 12, 2015

Paul Graham’s latest essay warns of losing time and focus to “corporate development” staffers out there sniffing for companies to buy. I agree with him on the corp dev side. But I wouldn’t want people to think that you should ignore potential buyers until “you want to sell your company right now.” It’s a question […]

Read the full article →

What is good product management?

November 24, 2014

I gave this talk to the CTO School in NYC the other week. The focus was on three key topics: what is good product management, how can the head of engineering be a good partner, and how to hire good PMs. Enjoy: What is good product management from Giff Constable

Read the full article →

The Virtual Thief (and what that means for privacy)

November 18, 2014

In 2006, we created a online marketplace for virtual goods in Second Life. Our competitor had a daunting network effect advantage, but we had no intention of fighting over the existing market. The real market in our eyes was consumers yet to come (they never really came, but that’s a different story). To drive traffic, […]

Read the full article →