The Most Important Thing About your v1.0 Product: “Product Surface Area”

Screen Shot 2014-06-29 at 8.03.09 PM

In this age of the “lean startups,” “agile development” and “minimum viable products,” the temptation for product designers is to get something out quickly and iterate with feedback.  So many great applications have been built that way. Early features are intended to grab an audience. The next step is to proceed to incremental development with a goal to deepen user engagement. It’s an attractive theory. What’s the point of inventing the future in a dark room and missing the market? Why not ship quickly and, rather than debate with your co-founders about which features matter, let the market decide.  A/B test your way to heaven.

If releasing a “minimum viable product” and then iterating quickly isn’t the key to a successful v1.0, then what is? The key is your “product surface area.”   Your product surface area is the broad collection of features that enable your potential user or customer to define you in their mind.  For example, by adding a broad set of use cases for publishing and sharing videos, YouTube’s product surface area created a mental space in its users that it was a “video sharing website.”  Even though a lot of the early use cases were centered on dating, it did not become a “dating website.”  The founders ensured that the product surface area was broad enough to go after the larger opportunity.  When you take your first product to market and pitch it to your first 10 prospects – you are holding your product and your business up to a mirror. You are asking the mirror, “How much do you love me; and what would it take for you to love me a little more?” The mirror will give you a totally rational answer. If you are medium build, slightly pudgy, wearing a yellow shirt and your hair is a disheveled, it might come back with, “you look pretty good.  Just lose a little weight, throw on a jacket to match the shirt and comb your hair.” But it certainly won’t come back and give you the path to become a fashion model. That’s the nature of your product surface area. Customers do a tremendous job giving you highly valuable incremental feedback but they can’t perceive value propositions that are well outside of the feature set. Therefore they can help make sure your stated value proposition improves but they can’t tell if a whole new bucket of features might be a lot more valuable. Given that you’re about to bet dozens of people, millions of dollars and your reputation on a product direction, the most important thing to get right in your v1.0 is the right depth and breadth of your product surface area. Go too broad and the execution will be terrible. The response you will get back will be a function of poor execution rather than a poor idea. Go too narrow and you run the risk that any feedback you get misses a much larger opportunity.

At BloomReach, we recently released SNAP (Search, Navigate and Personalize). It is a product with meaningful surface area. A “minimum viable product” approach would have suggested we just release a faster, better site-search product and iterate from there. If we did that, we’d get a lot of feedback on how to build better onsite search. Instead we believe that search technology needs to come together with navigation and personalization to deliver a new class of discovery on websites.

Building that combined stack took longer, but it ensured that the surface area was broad enough for us to claim a much larger market and make ongoing feature sequencing trade-offs across the larger surface area and that we received feedback across the full dimensions of the feature set. Picking your product surface area for v1.0 to balance timeline, scope, quality and opportunity is art, not science. While finding your optimal product surface area is important for consumer internet businesses, it’s critical for enterprise businesses.  Here’s why:

  • Feedback Cycles: Consumer products are about driving delight among millions of consumers. Enterprise products are about driving immense value among hundreds (maybe thousands) of companies. The sample size of feedback you get from your early enterprise customers will likely be much slower and much less representative of your broader market than in a consumer business (too few data points). Hoping to iterate on feedback simply takes too long.
  • Less Tolerance for Experimentation: Enterprises can’t tolerate radical experimentation.  They really need something that meets the test of someone working at a company justifying to their boss why they paid money for it.  As a result, you don’t get a lot of chances to take half-baked products to them and its important that the initial “surface area” be appropriate so you get the right feedback off of your v1.0, knowing there may not be multiple opportunities.
  • Reputation Matters More: When a consumer company’s experiment fails, it moves on to the next five experiments. In an enterprise-focused business, when a prospect is the one who tried your failed experiment, the only thing they can tell colleagues is “I tried it and it didn’t work.” Good luck selling to the prospect who received that report.

Consumer hardware devices may actually look a lot more like enterprise businesses in this regard. When Amazon released its phone, it seemed they were asking the Mirror – “if I give you the same full-featured phone as the iPhone with better 3D capabilities and a Prime Subscription thrown in, will you love me?” The Mirror will now speak, and I can assure you that if the phone has missed the mark, lightweight iteration won’t be the answer.

Next time you’re planning v1.0 – remember the Cinderella fairy tale, “Mirror Mirror on the wall, who’s the fairest of them all….”

Image from Surface Area by Jon Fife licensed under CC by 2.0

Software’s Valley of Death

Screen Shot 2014-06-08 at 10.36.26 PM

The early days of a software start-up are hard. You’ve got product that’s half-baked. Customer pilots that aren’t going well. Hiring engineers is brutal. You’ve got no one to sell your product. You’re running out of money and raising money requires a successful set of customer proof points. You know the playbook that consumer internet companies are following – don’t worry about monetization, just drive user value and grow. The money will come.

I’ve got bad news for you. If you’re in the software-as-a-service (SAAS) business, unless your business model is designed to ensure the money will come, the money won’t be there at the end of your journey.

In fact, few things are as important to the long-term viability of a SAAS business than the congruence of your price points with the market segment you are targeting. Lets do some simple math. Lets say you want to build a $100M recurring revenue SAAS business. There are really only two fundamental ways of doing it:

  1. Tackle the Enterprise: Sell 1000 companies @ least $100,000 per year deals with a market size of at least 5,000 companies
  2. Tackle the Small & Medium Business: Sell 100,000 companies @ least $1,000 per year deals with a market size of at least 500,000 companies

A company like Workday is a best in class example of No. 1, today worth $15 billion. They may have long sales cycles but it results in well over $400,000 per year deals. They went public with 325 customers and had a $135 million previous-year SAAS business. On the plus side: Deals are big; you can pay expensive sales teams their $250,000 on-target earnings; customers are sticky and contracts are typically multi-year so the business is enormously predictable. On the con side: I’m sure the business had greater customer-concentration in its early days than anyone liked; the early bookings must have been pretty lumpy and the sales/IT integration cycles must have been pretty long.

A company like Solarwinds is a terrific example of No. 2, today worth $3 billion. They built the business on small deals, typically less than $6,000 per transaction. Over 50,000 customers used the service at the time of the IPO to generate $62 million in revenue back in 2007. The model promotes that kind of velocity. It is inside-sales driven, with a downloadable version that reduces lead generation costs. Sales cycles are short because reps are often upselling a customer who has already downloaded the product. The market size is large in terms of number of companies, because almost any meaningful business can use the product. The deployment is quick. The pros are that when the flywheel is built, it can be a very cool business. The cons are that building that kind of high-velocity flywheel is super hard and not all market spaces lend themselves to this approach.

The valley of death is often in the middle. It’s the company with the long sales cycle (often because you are selling something that is pretty complicated to enterprises). You still require expensive sales people, have a limited market size and low price points (often $5,000 to $30,000 a year. ) Over time, the model simply doesn’t work. You can’t afford to sell or support large enterprises at these kinds of price points. There simply aren’t enough large enterprises to help you build a large business. Often it takes you a couple of years to discover this. You start off thinking “I will worry about building a great product first and worry about monetizing it later.” Then you convert your first customer into a paying one for $1,000 a month – thinking you can raise prices later once you’ve got customers to serve as references. Sometimes, you might even start with a higher price while targeting the Enterprise but start dropping it rapidly when you face competition, rather than fighting for value. You hire your first set of sales people and they keep closing low-price deals after long sales cycles. You assume that’s because you hired the wrong people or didn’t train the right people well. Your account-management costs are going through the roof and you’re wondering why you’re dealing with a bunch of custom requirements. You’re three years into your venture and you’ve landed in the valley of death. You’ve neither built the high price point, high value enterprise business nor the high velocity mid-market business. You’re simply stuck because neither outcome looks all that achievable.

The valley of death is totally avoidable, but only if you deal with it pretty early in the lifecycle of your start-up. Basically, you’ve got to pick your market segment in tandem with your product. You’re either enterprise, or you are mid-market (at least initially). You can’t be both. If you’ve picked the enterprise direction, somewhere in the first year – validate that someone will pay you at least $100,000 a year or iterate on your product till they do. If you’ve picked the mid-market direction, validate that you can address tens of thousands of companies and do so in a high-velocity manner early in your company’s evolution. If you are in the green zones, you’re set for greater things down the road. If you’re not, iterate till you are because the pain will be a lot worse down the road if you don’t. The great companies like Workday and Solarwinds design their entire business around their market segment – product, marketing, sales costs and competencies, account management and financial metrics.

A corollary strategy that many software entrepreneurs take is the “land and expand.” The idea is to sell a low price point, high-velocity product first – aiming to upsell later into a larger deal. That strategy (taken by companies like Dropbox or Box) can be a good one. But it doesn’t change the basic calculus of your market segment. Land and expand is a marketing/product strategy. If the net result is that you end up with small numbers of low-price-point customers, you’re screwed, whether you’ve landed and expanded or not.

Dealing with price can be particularly tough for product-oriented founders. The temptation is to keep iterating the product and deal with money later. Remember, this is B2B software and not the Consumer Internet space. Aim for the Enteprise or aim for the Small and Medium business, but at all costs, avoid the valley of death.

Image from Fire! by Eoghann Irving licensed under CC by 2.0

Democratizing your Culture

“Democracy is the worst form of government.  Except for all of those other forms that have been tried from time to time.”

– Winston Churchill

Screen Shot 2014-05-29 at 9.02.52 PM

Culture is the center of the start-up.  It is the pulse.  It is the soul.  It is the binding that gets teams through hard times.  It is the intoxicant that enables teams to share the highs with each other – making them even higher.

At BloomReach we have our own core values, written in February 2009, before my cofounder Ashu or I had raised any money, figured out our product, determined what market we were going after or hired our first person.  That’s how important it was to us.

The most interesting question around culture is not, “What is your culture?”  It is, “How do you make your culture real?”  Thousands of enterprises have come before your start-up or growing business.  Most of them have terrific culture and values documents.  But they are not real.  They are not authentic to the team or to the business; and they don’t manifest in ways that are practical to the work lives of employees.  The pseudo-cultures think that the key to great culture is a lot of parties, free food or cool schwag.  It’s not.

The key to making your culture real is to democratize it.  The aspirational goal of your culture should be to make it genuinely co-owned by every member of your team.  That can be counter-intuitive because so much about companies is top down – the titles, the strategic direction, the planning, the compensation and the decision-making.  But culture needs to be built bottom-up.  It can’t be the responsibility of the HR team or even the CEO.

It has to be an organism that grows and shrinks with the ebb and flow of your business and the personalities of your team.  It has to be truly democratic.  So how do you democratize your culture?  A democratic culture draws from the pillars of a democratic society.  It starts with citizenship.

Here are 10 ways we do it:

  1. The BloomReach Citizenship Document:  We maintain a document of dozens of initiatives – broken up between “Fun” and “Company Building” in a Google doc spreadsheet.  Each initiative has an owner and a team.  Initiatives can be “improve remote office communication,” “organize fitness activities,” “organize peer awards” (more on that later) or “create volunteer events.”  No one approves these initiatives.  The only ask of a new BloomReacher is that he or she do something that goes beyond their job description to contribute back to the company.  Many of their contributions are memorialized in the BloomReach citizenship document.  It serves as a repository of our “democratic traditions.”
  2. Hackathons and Business Challenges:  We’ve had a tradition at BloomReach of either running hackathons (24-hour efforts to build and ship something cool) or running the BloomReach Challenge (where the entire company is divided into cross-functional teams to dream up business and product ideas). Hackathons and business challenges encourage people who never get to work together and often don’t know each other to get creative and get to know one another.  It also reinforces that just as the rewards belong to all of us, the challenges belong to all of us.  Eighty out of 170 people participated in our last hackathon and several new product features came out of it, promoting the energy and creativity that are “must-haves” in any start-up.
  3. The Open Floor Plan: Democratizing your culture starts with open communication among teams.  Just as in any democracy, there must be freedom of expression. The open floor plan creates energy, expression and sends a clear statement that the environment values everyone equally.  Yes, it can lead to some loss of individual productivity, but the cultural gain across the organization outweighs all of that.
  4. One Set of Rules:  Are you flying first class and expecting others to travel coach? Is an exec allowed to spend exorbitant amounts of money on a team dinner that a team itself would not be allowed to spend on its own?  All of these send clear signals – that the culture does not value the contributions of every member of the team equally.
  5. Peer Awards: Peer awards are fantastic.  They reinforce that the highest honor anyone can receive is an unsolicited award from a teammate.  For us, peer awards are a big deal.  Anyone can give them as a thank you to a colleague, along with a $150 gift certificate.  It comes with an Oscar-like thank you speech and creates the essential quality of a democratic culture –the willingness to go to great lengths to help a teammate in need.
  6. Bonuses on Company Performance:  A key to democratizing your culture is the idea that “we rise and fall as one company.”   And there is no credibility to that claim if the compensation of individuals is meaningfully at odds with how the company performs.  Of course, one should find mechanisms to recognize extraordinary performance but a democratic culture means democratic compensation features.  Because base compensation and equity are often variant on time of joining and role, a target bonus percent purely based on overall Company performance helps equalize the compensation mix.
  7. Independent Ownership of Decisions: A pillar of the democratic culture is the idea that each individual is an adult, capable of making good decisions as an owner of the business.  At BloomReach, we borrow from the Netflix model of limiting “policies” – no vacation policies, no expense policies and no over-legislation of behavior.
  8. Self-Scoring of OKRs (Outcomes and Key Results): Nothing is more typically top-down than individual performance.  On the other hand, self-scoring of team OKRs enables teams to (in a non compensation-impacting way) score themselves in public and share those scores openly.  It holds each team accountable to each other.
  9. 360 Feedback: We try to keep our performance-feedback process simple.  However, one of the key changes we’ve made is to incorporate 360 feedback (particularly, feedback from peers).  Top down feedback reinforces that the only thing that matters is pleasing your boss.  In truth, high-impact initiatives involve teams collaborating with each other fruitfully.
  10. Hanging out together:  You can’t force friendships.  But you can put people in a position to get to know each other genuinely.  We do that with a lot of company-sponsored initiatives like soccer teams, volunteer events and running relays.  I’m proud of the number of BloomReachers that hang out with each other.  Social relationships build camaraderie.  And camaraderie creates loyalties that go beyond loyalty to the company. (Ok, there are some parties involved.)

I believe that a democratic culture solves the single most difficult conundrum of growth: scaling while maintaining extraordinary commitment.  We have a long way to go to continue to evolve what being a BloomReacher means.  I am sure, however, that the answer to every impending decision lies somewhere in the historical tradition of the only system that has ever really worked: democracy.

Photo courtesy of “DEMOCRACY” by SUXSIEQ under CC BY 2.0

Hiring a Great Executive Team


One of a CEO’s hardest and most important jobs is building an executive team.  Have a great executive team? You can power through the darkest of hours. Have a dysfunctional, less-than competent or highly unbalanced executive team and you’re in for a world of hurt.

Don’t Rely on a Retained Search Firm: I’ve never been very successful in hiring execs through retained search firms. (We are 0-3 on searches done through retained search firms at BloomReach). That could be because our standards are insanely high. It could be because we are super-sensitive to “fit” – something that is very hard for a retained search firm to get right. Or it could be because retained search firms focus primarily on people in the market, who may by definition be lower quality candidates. I’m not saying you should not hire an outside firm but it is essential that you use your network to identify the absolute best candidates also– then sell them like hell independent of an outside firm. This will result in a smaller pipeline, but a much higher quality one.  Remember, it is your responsibility to find the best candidate.  Not the firm’s.

Know when to bet and when not to: The most common comparison point for CEOs evaluating candidates is between the “experienced hire” and the “up-and-comer.” Some companies believe in one philosophy or the other. I think the answer is situational. We have members of the BloomReach executive team for whom this is very analogous to what they’ve done before. We have members that were total bets that worked out. We have people we have promoted from within. We have people that we have added recently and those that have been with us for 4-plus years. A good rule of thumb for me is, if you believe that the job requirements are highly creative, unique to your business, an area you want to “invent” and spend meaningful founder-time on, then hire the up-and-comer or grow the person internally. They will be more open minded, have less baggage and will be more able to think out of the box.  If you believe you have more to learn from the best-in-class companies in your industry for that function, or if you as a CEO/founder want to spend a little less time on that function, hire for experience.

Hire for Raw Smarts and Intense Motivation: Most start-up journeys involve navigating uncharted waters. I have always made mistakes when I have hired based on resume. Every candidate is only as good as the situation they are placed in. And in most start-ups, that situation relies on versatility, intelligence and sheer (street) smarts. The ivory-tower candidates are better served working in larger corporations or universities than in your chaotic start-ups. Motivation counts as much as intelligence. The ideal candidate is one for whom their success at your company is a defining career moment; where their passion and commitment approximates your own.

Fit Matters: Most unsuccessful executive team members fail because they don’t share the culture or work collaboratively with others on the team. Ask yourself the following question: Would the new exec team member I’m hiring add to or detract from the harmony of the team. If that answer is not obvious, don’t hire them. Run the interview process through your executive team and take the feedback seriously, don’t pre-judge the outcomes.

Master of One Function, Jack of All Others: The standard for our executive team is a clear one. Team members must be masters of the function they are leading (engineering, product, sales, marketing etc.). They must be able to get down to the level of an individual contributor on their team (and answer your hard, detailed questions) and rise to scale it at the same time. Without that, they can’t do the job. They must also be able to add meaningful value to every (controversial) conversation around the exec table. That means Administrative/Finance people need to understand the business. Product people need to understand customers. Customer-oriented individuals need to understand technology. And everyone needs to understand the culture and the business drivers.

Perhaps as valuable are the qualities that don’t matter (except as they influence the above): You are not looking for your best friend; you are not looking for the best resumes; you are not looking for the “silver bullet” whose magical arrival fixes the business; you are not looking for the best financial deal on your hire; and you are not looking to have it be primarily the group of people that helped you get off the ground (i.e. your earliest employees) since the needs of the business will certainly evolve.

In the two years since we have assembled our core executive team, I can say clearly that this is the best senior team I have ever worked with. And I sleep a lot better at night knowing that.

Photo courtesy of “Dragon*Con 2013: JLA vs Avengers Shoot” by Pat Loika under CC BY 2.0

Be Splunk, Not Box


If competition matters, you haven’t innovated in your market definition.

The companies I admire most stand in their own image.   They can’t easily be defined by any definition of a market that Gartner or Forrester have defined and written reports about. They chart a clear mission or goal – and that goal substantively redefines a market or a product category. Because their world-view is so unique, it ultimately enables them to stand on their own, much less concerned about competitors than about adoption of that world-view. Lets take two success stories in the Big Data / Software world as examples: Splunk and Box. Both are highly successful companies with likely multi-billion-dollar market capitalizations.

Here’s what Splunk’s S-1 said about its competitive environment at the time of its IPO:

“We compete against a variety of large software vendors and smaller specialized companies, open source projects and custom development efforts, which provide solutions in the specific markets we address. Our principal competitors include:

  • IT departments of potential customers which have undertaken custom software development efforts to analyze and manage their machine data.
  • Security,systems management and other IT vendors,including BMC Software,CA,HP,IBM,Intel,Microsoft,Quest Software, TIBCO and VMware.
  • Web analytics vendors, including Adobe Systems, Google, IBM and Webtrends.
  • Business intelligence vendors, including IBM, Oracle and SAP.
  • Companies targeting the big data market by commercializing open source software, such as the various Hadoop distributions and NoSQL data stores.
  • Small-specialized vendors, which provide complementary or competitive solutions in enterprise data analytics, data warehousing and big data technologies that may compete with our software.”

Huh? They compete with everyone from custom software to web analytics to business intelligence to security to systems management vendors?

Here’s what Box’s S-1 said about its competitive environment:

“The market for cloud-based Enterprise Content Collaboration services is fragmented, rapidly evolving and highly competitive, with relatively low barriers to entry for certain applications and services. Many of our competitors and potential competitors are larger and have greater name recognition, much longer operating histories, larger marketing budgets and significantly greater resources than we do. Our competitors include Citrix, Dropbox, EMC, Google, and Microsoft.”

Rough. All they have to do to succeed is win against a rapidly innovating start-up with a disruptive adoption model (Dropbox), and three of the largest software businesses in the world (EMC, Google, Microsoft).

Splunk had raised $40 million by the time of its IPO. Box had raised $400M. Their respective business models had a lot to do with the capital requirements but the primary culprit was the competitive environment for each of the two companies. Box simply needs ungodly amounts of money to achieve dominance in its market versus Splunk’s more modest needs. Why?

Because Box did not innovate in its definition of the market, while Splunk did. Both companies innovated in their early products – they would not have gained meaningful adoption otherwise. But product innovation isn’t nearly enough. It’s market definition innovation that is truly sustainable. In Box’s case – it was clear they were selling a collaboration and file storage service. That market existed long before Box got there and will exist long after. Even SAAS versions of such a service existed. In Splunk’s case no one was very clear what they were – were they a log management company? Were they a search company (a lot of the early innovation was search on log files)? Were they a systems management company (a lot of early use cases were around keeping systems up and running)? By having a murky market definition, Splunk could compete against legacy vendors in all of those categories because no one could neatly put them in a box (no pun intended).

This meant that venture capitalists funded fewer competitors and each of the major players in the space (large security or log management or search or systems management companies) all just walked away. Why? Because none of them thought it was their market to compete in. Of course, as the category evolved, large entrants started to pay attention. However, it’s all too late. It’s not enough for a competitor to compete by simply building a better product, or throwing capital at the problem, or arriving with better distribution. They have to learn completely new businesses. Security companies don’t have systems management DNA. Systems management companies don’t have search DNA. And Search companies don’t get selling into IT. What a terrific competitive moat!

Don’t get me wrong – you can compete with better execution in a highly competitive market. Just prepare to raise several hundred million dollars and be an execution machine. The hardest thing about innovating on market definition is that you don’t get to have a well-defined (and easily communicable) total addressable market, or TAM. Your products and your proposition define your TAM. Your TAM does not define your products.

At BloomReach, we think a lot about market definition.   And I love that no one can put us in a box. Are we an SEO company? A big data company? A site search company? A personalization company? An e-commerce enabler? A SAAS company? World-views are powerful things. And there is nothing like standing alone in your world-view and being right.

The Star Performer Hype Cycle

When it comes to building a team, venture capitalists and experienced leaders spout a classic refrain: “hire slow, fire fast.”  The logic makes sense at one level.  You only want to hire A players – in skills, cultural fit and motivation.  And often within the first 30 days, you find that new hires are exactly who they ultimately will be. Therefore, take your time finding the fit and if you determine that the person is not right – fire them.

But here’s why it often makes sense to hire fast:

  • The market for great talent is hot… as hot as it has been in the Valley in years.  Decisiveness counts for something in identifying the right candidate.
  • In many cases, more interviews just create confusion. Consensus on a candidate in a candid, aggressive culture almost never happens.
  • Interviews are fundamentally flawed.  The probability of a better hire with more data degrades fast when the macro issue is that interviews yield highly imperfect results.

Firing fast is the more dangerous decision.  Firing fast has cultural implications – it creates a lot of turnover, it sets initiatives back, it often creates politics and it creates a murmur of “why did Person X leave” (assuming one does not disclose publicly who leaves voluntarily and who doesn’t)?  If the hiring bar is really high (as it is at BloomReach) – firing fast creates a simple message for the team: “You are only as good as what you do in the first 90 days?”


We’ve had plenty of amazing people who were put in the wrong role, struggled to understand how decisions are made initially or were slower (but ultimately more productive) learners.  And culture isn’t something that someone gets overnight.  It takes time and osmosis to build relationships.  I have found that most hires operating in a high-performance organization go through their own version of the Gartner Hype Cycle. (Lets call it the Star Performer Hype Cycle.)  It looks a lot like this:

hire fire


In the Star Performer Hype Cycle – the trigger is the hire and often in the beginning the star is a breath of fresh air.  They bring in new ideas, new contacts and a new personality into the mix.  It often feels like they have the silver bullet to all of your problems.

Until they don’t. That’s when the peak of inflated expectations comes crashing down into the trough of disillusionment.  That’s when you feel like they don’t sell enough deals or get your technical architecture or even understand the most rudimentary concepts of the business.

For stars, at some point something clicks.  They know how to execute in your chaotic environment.  They sell one deal and that helps them sell the next.  Confidence breeds confidence.  And the virtuous cycle grows.  Extricating your new star from the trough of disillusionment can sometimes take an intervention from their manager: a refocus on priorities, a clarification of role, a therapist to listen to their struggles.  I’ve taken a lot of BloomReachers through this trough and out to the high performance category.

Fire fast and risk falling into the thrall of the “grass is greener syndrome” – betting you can replace him or her with instant greatness.  Remember, your star performer may be one quarter away from greatness.