30 West 3rd

Very Early Stage Technology Investing

Archive for the ‘Business Application Software’ Category

What “great team” really means (Part II – Software/Business Applications)

leave a comment »

We covered building the unique elements of a bioscience team in the last post.  Now we will look at building a great software team, with some thoughts on what that means in traditional software products, business applications and consumer digital companies.  Most of the digital opportunities I see don’t emerge from research institutions, so we will focus on software that emerges from the general community of talent in the world.

For a horizontal software application – think of databases, utilities, b/i tools, etc. – that are primarily sold to IT buyers for IT users, the founder is typically a practitioner who sees some gap in the always creatively destructive software market and fills the breach.  This happens constantly, the infrastructure of what we see in business applications, is constantly being rebuilt using new tools.  These evolutions happen around every 10 years it seems, but there are segments of the infrastructure world that change under the covers while the user experience doesn’t change.  Buzz word though it may be, the movement to cloud computing may be the most disruptive yet, as it is obliterating the cost structure of a lot of existing companies and markets, while creating interesting opportunities for others.  There are large software companies in the enterprise space that are under assault from Amazon just as much as the traditional book business.  A good founder in this space is both a good practitioner, but may not a great one, but has the ability to look at the impact of changes in the complex world of IT infrastructure and develop something quickly that, while not need today, will be needed by everyone in a few years.

A great example of that in my career was Gaurav Dhillon, founder of Informatica in the 90’s and now the founder of SnapLogic.  I met Gaurav when Informatica was still in fresh in the market.  He explained how he had come up with the idea for a better way to move data – it was batch oriented at the time – while working for a large company in Europe, but they didn’t think it was that interesting.  Gaurav knew that the data warehouse boom was coming and that people would demand a better way to move data from transaction systems to large data warehouses.  And because he knew it would happen he founded Informatica and willed the company to become an IPO and continued relevance today.  He built a great team, surrounded himself with smart investors and simply out-performed the field.  Today Gaurav is back with a relatively new start-up that focuses on how to move data with cloud infrastructure.  The need is the same, but the game is changing and he knows it.  I wouldn’t bet against him and this blog post by Ben Horowitz perfectly describes what a founder looks and acts like.  I love this quote: “Although, Informatica is considered a great success, it isn’t a great success for Gaurav, because he deeply believes that both the idea and his ability to execute it exceed the outcome that he achieved.”  And if you read the entire post you’ll see both that Gaurav made a lot of money at Informatica, and that money does not drive him.  Most great founders are shocked to be rich when they get there.  They almost forget that is a by-product of success.  In the application software market you need a bit of a different kind of obsessed practitioner, more of a business leader in a functional area, but they need to have the same insight around how software responds to business needs.  Both Tom Siebel and Marc Benioff were executives at Oracle with sales experience.  They each saw the need for CRM, and one eventually displaced the other by deploying a new type of infrastructure.

The founder of the great software company is almost always a great product evangelist and will drive the first sales.  But early sales don’t mean much unless they can be repeated by less visionary, less willful sales people or channel partners.  In the end one person can only sell so much software on their own, so the great founder now needs a great sales leader.  I have described these characteristics in a previous post, and I don’t really think much has changed.  Read it again for the key qualities, but you cannot compromise on this position.  Too many CEOs mistake their ability to sell with willfulness as a gift that can infect others.  The opposite is more likely true.  Sales success comes from simplifying the message so that any competent, trained sales rep can deliver it.  You shouldn’t need virtuoso skills to sell a compelling product.  You need process, good packaging and pricing, and great competitive materials and training.  A good sales leader will have carried a number bigger than your 2 year plan, know the first five people he/she is going to hire, and have a Rolodex (there’s a Brand name that endures longer than the product itself) of qualified buyers that they will have already talked to before they come on board.

Some companies need a CTO, but it’s not always an early hire.  The mistake is often to put the founder’s coder friend in that role, but that makes sense only if that person is a visionary, good in a market facing role, and really owns the product vision.  And I don’t know that too many business application companies need a CTO.  And having a CTO does not obviate the need for an engineering lead.  Someone with product development experience who is willing to be held accountable to a schedule is first priority, then figure out if you need a glamour CTO.  The most underappreciated position in any software company is product manager.  You need a product champion that manages the natural tension between market need (tractable, infinitely broad) and technical team output (overly optimistic, sometimes too creative).  Software is just too easy to write today, and it is very easy to create too much of the wrong thing and little of the right thing in an application, website or other product.  The product manager must be technically conversant, maybe a developer who hates to code but is good w/ users, but not impressed by just more output.

The team is rounded out by someone that can drive awareness and lead generation for the company, but great early stage sales leaders and their teams are usually not terribly dependent on marketing.  I think it is often better to invert the concept of lead generation in the beginning.  One of our companies has about 200 target customer to tackle this year.  The buyer is easily identified, the readiness of each account to buy is probably easily discerned in a first call/visit, and then you can move on.  A similar approach, maybe easier, works when you are building a company through indirect sales.  The best distributors can be identified and pursued.  Marketing is an important task, but a great company in the hands of great sales/biz dev people need less lead generation and more competitive information, case studies, and other things that bring credibility to the transaction.

Finally, you need a great transactional lawyer and a great CFO/VP of Finance with industry experience.  This may not be a full-time need in the seed stage, probably not.  But you do need a robust revenue and cash flow model, industry standard contracts and transaction advice, and all the right employment agreements, comp plans and other operational tools of the business.  This is a lot to assemble, and we try to add what is missing, augment with proven p/t professionals, and then scale the team.  I do think a seed stage company can’t make much progress with three good legs to the stool.  In a software company that’s the founder, sales lead and VP of Engineering.  In Part III we’ll talk about how a great team really works together, how they act, and the most important characteristic they must share.


Simple designs win Web software fans…

with one comment

User interface design has lagged other areas of computing since the 1950s. Early computers were complex, expert-only devices.  Interface methods ranged from switches and dials to key punch cards.  Only specially trained personnel were able to interact directly with the machines.

The emergence of keyboard data terminals in the late 1970s allowed expert programmers to type in commands.  Finally, in the late 1970s, specialized end-user applications emerged.  This event was revolutionary to the industry, requiring accountants, actuaries, and other business end users to interact directly with a computer rather than pass instructions through a programmer or specially trained data-entry clerk.

These early end-users were the first to complain about interface design.  The interface at the time was a simple monochrome screen with characters.  No graphical elements or icons to build a rich user experience.  No mouse to navigate.  No multiple windows to support multiple tasks.  In this world, computing tasks were highly linear, with step-wise data entry and batch reporting as the principal end-user functions.

I entered the industry in the late 1980s and lived through the graphical user interface (GUI) transition in computing.  The introduction of graphical, pixellated, multicolored screens ushered in a new age of application possibilities for the industry.  Unfortunately, application design methods and standards were divorced from how people thought and how tasks were accomplished.  There also was a steep learning curve for the industry on fonts, colors, contrast and other visual elements of design.  About the time application design had evolved to a reasonable level of competence, the Internet came along and reset expectations again.

The principles of good interface and application design that were well established when Mosaic was first introduced created outsized expectations for end-users about what could be done online.  The PC-based client-server computing environment was a complex system that supported far more intricate application elements than the Internet could at first deliver.  Even today most online applications that gather and manage data are primitive in structure and design compared to their client-server ancestors.  Moreover, Ajax and other client-side Web interface elements represent nothing more than a continual thinning of the client-side of computing.  This is true of mobile applications as well.

The real revolution driven by the Internet is the proliferation of training-free, intuitively obvious applications that address mass audiences.  Even when user interface design reached a reasonable level of proficiency in the mid-90s, the typical user was a business employee taking an order, working in a call center, running a financial report, or hiring a new employee.  Each user went through specific training on how to use each application.  User-support specialists stood by to help everyone use each business system. The Web changed that completely.  All users now have consumer-based expectations for application design.

One simple example is travel.  If you are old enough to remember travel before the Internet, you will remember that average individual travelers had no direct computer interaction with reservation systems.  Travel agents and airline/hotel/rental agency reservation agents were trained in a handful of complex applications to make reservations.  The main system, SABRE, still lives behind most travel Web sites in some form.

The Internet allowed the casual consumer to make reservations directly.  To accomplish this, the travel industry had to design user interface elements that were immediately clear to the end user, who had no special training or previous exposure to the system.  This is why Web site design revolves almost entirely around simplicity, obviousness of intent, and field-based data entry.

Today the formerly labor intensive task of taking reservations in the travel industry has been entirely outsourced to the consumer.  Making a reservation using any of the major travel sites is remarkably similar in each case, and few novelties have been introduced.  In fact, a travel site missing a pop-up calendar or violating the basic and well-known steps to complete a reservation would likely fail.  Incremental improvements to the process are allowed, but keeping things simple and task-oriented is the first rule of Web application design.

These simple principles have cascaded into business applications.  End users are trained to use computing resources first on the Web and next on their phones.  Business applications that present a more complex interface design are not likely to meet with broad acceptance.  The evolution of Web design, with a cult-like focus on customer intimacy and task focus, means less innovation in bells and whistles and more innovation in simple metaphors of usage.  It is critical for early-stage Web companies to get this right from the outset.  Tight functional circles defined by clear design elements will attract consumers or business users.  Forget simplicity and complexity will seal your fate.

Written by Mike Venerable

October 13, 2009 at 8:50 am

Great Sales Advice

leave a comment »

A friend sent me this post from the NY Times BITS blog last week.  It contains some solid advice on software sales execution from Sohaid Abbassi of Informatica.  Informatica was blessed in its early years with some great sales leadership, including Clive Harrison (most recently at Exeros) and Mark Burton (ran MySQL prior to the Sun acquisition).  The company had a great pre-bubble run of growth under founding CEO Gaurav Dhillon.  Back then they won consistently on sales execution, and seem to be doing that again, more than 10 years later.  And of all the BI/DW companies of that vintage, they appear to be the only significant stand-alone survivor.

Now, to the advice.  During the downturn the company focused on near-term deals first, and tracked account progress weekly.  They focused on where purchase approval paths were clear and understood.  This was done in response to the current economic contraction, but the advice and approach would be a good daily operating model for market-entry software companies.  Most early stage companies struggle with sales and lack the experienced leadership or discipline required to define a clear sales process.  Qualification discipline is typically lacking, and is almost non-existent if the market path is indirect.

Good sales process metrics should be defined first using an analogous example from an existing company.  Hire at least one successful and qualified sales professional to work prospects as early as possible.  New products are first sold on the personal relationships and will of the founder(s).  That is great, but it will not scale.  At some point the product must be sold by sales people or distribution partners who can articulate a value proposition, qualify opportunities, and close deals without hand-holding.  Recognize also that early buyers are typically more vested in the product than later ones.  There is often a mutual feeling of creation and excitement around new products among the first 3-5 customers.  Normal purchasing behavior is ignored, competitors kept out, and everyone is working towards a successful first transaction.  Not so with later customers.  To succeed in scaling revenue the early stage company should focus on process discipline, hiring proven sales talent, and listening to the market feedback they provide.

Written by Mike Venerable

September 14, 2009 at 12:54 pm