David Cummings on Startups

Blog

  • Setting the Corporate Culture Tone

    Most entrepreneurs, especially when they are starting out, underestimate the power of a strong corporate culture. In my experience, the corporate culture is immediately evident upon entering an office and walking around. How can you tell? Here are some simple examples:

    • Awesome office (fun, cool, and inspirational)
    • MacBooks and iMacs on desks
    • Dual monitors everywhere
    • Unlimited drinks and food
    • People that truly want to be there

    My recommendation is to read about the Zappos corporate culture as well as books by Jim Collins and Patrick Lencioni.

  • Product Death by Not Enough Cuts

    In the software world, death by a thousand cuts comes from trying to make the product do all things for all people. I want to bring up another phenomenon not talked about enough: death by not enough cuts. This happens when an entrepreneur builds a product and gets it in the hands of a few customers, only then to keep customizing it further for those too few customers. In addition, it is often combined with not being opinionated enough about the product functionality, so a few companies drive the product road map.

    The next thing the entrepreneur finds out is that they have a deep and rich product, but not a big enough market to be successful. Oh, and the product is so large now that adding functionality takes significantly longer than it used to take — big problems lie ahead. Here’s what I recommend:

    • Get the product into the hands of as many prospects as possible
    • Launch early and often
    • Always ask yourself if the feature request fits into your vision for the next three years (be opinionated!)
    • After you’ve asked if it fits in with the vision, ask yourself if it is applicable to 80% of your user (e.g. will they love to use it)
    • Every 20 customers you sign up, raise the price until you can’t raise it anymore

    What else? Have you seen this happen? What would you recommend?

  • Iterate or Die – Part 5

    Once we had a few sales under our belt, we began to notice a trend in the higher education vertical. It was the one industry in which we were able to generate the most leads using PPC ads. Additionally, we found that our application was uniquely suited to the types of challenges higher ed clients were looking to solve. It was an ideal match, and one we had never anticipated in our early days as a company.

    Colleges and universities typically have a collection of independent websites with little or no consistency. As you might imagine, the development of these sites happened organically and in piecemeal fashion. Different technologies like PHP, Classic ASP, ASP.NET, and ColdFusion were used to power the dynamic portions, while plain HTML was used for more static sections. Our software can handle each of these situations by publishing files as well as content to remote databases. This flexibility has been a key differentiator for our product, especially for our higher ed clients.

    Our product’s special sauce is its ability to manage multiple websites, all of which live on multiple servers and use different operating systems, from one single product instance (just think of all the independent websites at Georgia Tech). It’s a difficult problem, but our product offers the capabilities needed to solve it. In addition, we had a simple per-CPU pricing model with unlimited sites, users, groups, and content combined with a focus on XML (before XML really hit the mainstream).

    I wish I could say we planned it that way, but we didn’t. The reason we could publish to different servers and support all the major programming languages was because of the goal we had set out with in our first (albeit unsuccessful) SaaS CMS: support all small business shared hosting accounts. With our SaaS CMS focused on small businesses, the only way to get content to their server was through FTP or SFTP. There weren’t any other options.

    With our new mid-market CMS, we set out to provide all the benefits of a dynamic, database-driven website with the performance and flexibility of publishing flat files. It turned out, unbeknownst to us for some time, that this arrangement was perfect for higher education. We started signing more and more higher ed clients, allowing us to build a portfolio of reference customers in a specific vertical — an important step in laying the foundation for future growth. With these valuable customer references in place, it was time to grow a serious business around our product. At this point, I had been running the business on my own for over four years and was barely scraping by; but it wasn’t until then, at the start of 2005, that I was finally certain we were on to something special.

    The new strategy was pretty simple — cold call all 4,160 two-year, four-year, public, and private colleges and universities in the U.S. and Canada.  We focused on calling people with the following job titles:

    • Webmaster
    • Director/VP of IT
    • Director of University Relations
    • Web Manager
    • Communications Director

    I had three full-time sales people cold-calling and setting up web demos for me. My role was to be both the tech-savvy sales engineer and the passionate product manager that gave the demo. It worked beautifully. Sales tripled in 2005 and more than doubled in 2006. We had finally hit our stride.

    Now, in 2009, we have over 120 colleges and universities as clients (including Emory, GA Tech’s Business School, Clemson, and Duke), making us one of the top higher education CMS vendors in the world. At the end of the day, achieving success came down to the following:

    • Neverending determination to succeed
    • Making decisions quickly and figuring out what works and doesn’t work
    • Being passionate about the product and the market opportunity

    Building a company is an amazing journey that is worth every minute. Even though you’ll frequently take many different turns and struggle, in hindsight these are a major factor in success. In addition, being able to iterate and learn quickly is one of the most important traits that a management team can have. Good luck!

  • Iterate or Die – Part 4

    After we had built an award-winning product and accrued several years of domain experience, we embarked on the next major iteration of the business: sales and marketing. I knew the technology aspects of my business inside and out, but my knowledge of B2B sales and marketing left much to be desired. After a bit of Googling, I decided to focus on three strategies:

    • Cold calling
    • Partnerships
    • Pay-per-click ads

    Cold Calling

    The first idea for cold calling was to buy a list of all the Chief Information Officers  in the Southeast with revenues between $100 million and $1 billion, a grouping often referred to as mid-sized companies.  We targeted the Southeast because we figured we would have the most success if we could meet with our prospects face to face. That’s another benefit of being based in Atlanta — as a regional transport hub, you are within reach of many other large regional cities, not to mention connected to a multitude of flight routes from the world’s busiest airport.

    Once we had a list of target prospects, we hired a slew of interns from nearby Emory University to cold call 1,000 organizations. The end result?  A measly four appointments. Our cold calling efforts, while valiant, suffered from a lack of the following:

    • Compelling value proposition
    • Referenceable customers we could name-drop
    • Product or company name recognition in the market

    Cold calling would eventually become one of our most effective strategies, but it took us a solid year to determine where we should focus our efforts.

    Partnerships

    I’d long assumed that finding implementation partners and resellers was a logical strategy to build my business. Potential partners, such as interactive agencies and ISVs, would provide the services and we’d provide the product. It would be an ideal win-win situation.

    Unfortunately, it never brought the results we’d envisioned.

    We continually worked over the course of several years to forge partnerships with ten companies. Those ten partnerships together resulted in just five total sales. That’s right — very few partners would officially sign on, and even fewer relationships resulted in actual revenue gains. It took me a long time to understand that introducing a mid-market CMS to a client would in turn reduce the amount of money that agency could bill for fees.

    Agencies operate in a time and materials model, and tens of thousands of dollars in CMS costs would come right out of the same client budget as hourly fees. From the agency’s perspective, building a custom client solution, even if it cost more and accomplished less than an off-the-shelf CMS, was the right thing to do for their business model. It was a hard lesson to learn.

    Pay-Per-Click Ads

    Pay-per-click (PPC) ads are the sponsored ads that show up alongside search results in Google and other search engines. When we did our first PPC campaign in late 2003, it was much more affordable and cost-effective compared to today. Quite by accident and much to our delight, PPC ads, combined with landing pages, generated leads in a variety of industries. After reviewing these leads, we methodically followed up on the promising ones and began moving prospects through the sales process. By the end of 2004, we had signed clients representing each of the following verticals:

    • Management consulting (1, from cold call)
    • Healthcare (1, from cold call)
    • Utilities (1, from partner)
    • Hospitality (1, from partner)
    • Technology (1, from PPC)
    • Higher education (2, from PPC)

    Head on over to Iterate or Die – Part 5 to find out what single industry we had slightly more success with and how we focused on that industry to become a market leader.

  • Iterate or Die – Part 3

    After initially building a small business SaaS product and subsequently licensing it to a larger company while moving the company to Atlanta, we had finally settled into developing what would ultimately become our successful flagship product: a mid-market web content management system. Of course, we had no idea at the time if this iteration would even be successful. One thing we were sure of, however, is that robust website management applications were complex and there was not yet a clear winner in the mid-market segment.

    The pre-paid royalties we were receiving from our licensing arrangement gave us the luxury to spend a year focused solely on building the application.  In his book Four Steps to the Epiphany, Steve Blank argues for customer-driven development, which is exactly what it sounds like: Customers tell you what they want, and then you build the product accordingly. Because we didn’t know any better at the time, we weren’t following this model.  However, we were aware of the sorts of things that our prior customers had wanted in a product. We also knew what functionality we wanted to manage our own website. I’m a big believer in eating your own dog food.

    Before long, we were ready to launch our new content management system. It debuted on April 15, 2003 at the Internet World trade show in San Jose, CA, where it promptly won the Best of Show award. The outlook for our new product was bright. Unfortunately, we soon learned that a promising product launch doesn’t always equal stellar sales.

    After working non-stop trying to sell the new application, we only managed to sell one license by the end of 2003. Mind you, landing a deal for a single $30,000 server license sure felt like more of an accomplishment than selling multiple lower-priced licenses for our previous product had. But selling only one server license over the course of six months was discouraging.

    After launching the first version of a great product, winning a prestigious award, and signing our first full-price client (we had given several licenses away for free to get early users with North Highland in Atlanta being our flagship reference — thanks Monica!), the company’s next major iteration was learning how to sell and market the product more effectively. Lead generation was the first area we focused on and continuously iterated. We tried these different tactics:

    • Cold calling IT directors
    • Channel development through interactive agency partners
    • Pay-per-click ads on Google

    Head on over to Iterate or Die – Part 4 to learn what worked and what didn’t work with our sales and marketing efforts.

  • Iterate or Die – Part 2

    In Part 1, I discussed the first iteration of the Hannon Hill business model, wherein we built a SaaS web content management system and launched it in the summer of 2001.  Initially, we charged $30 per month per website, but after realizing that there wasn’t a reachable market large enough to ensure our success, we quickly changed our approach to offering an installed version of the same product. To our surprise and elation, we sold our first $1,000 license by the end of August.  With this auspicious start, we were sure we had found a path to growth and profitability.

    Over the next 24 months, representing the life of the application, we sold 25 licenses. Not surprisingly, it turned out that selling just over a thousand dollars’ worth of software per month didn’t make for a sustainable business model. Our big break did finally come, though, in December of 2001 at the Internet World trade show in New York City and was a direct result of changing our gameplan to pursue an installed software model.

    At the trade show, we serendipitously met another, much larger software company that was explicitly looking for a content management system to sell to their installed client base of over one million licensees. We had several rounds of discussions with their management team at the show and eventually consummated a deal in Spring 2002 to license our product to them to be sold under their own brand name.  They would pay us a minimum monthly payment for the first twelve months along with a royalty for each license sold once the pre-paid amount was cleared.  We had our first big break!

    At the time, I had just finished my degree at Duke in Durham, North Carolina. With a guaranteed revenue stream in place by the end of 2002, I moved the company to Atlanta, Georgia in August 2002. Why Atlanta? Growing up in Tallahassee, Florida, Atlanta was always the big city we’d go to for a Braves game or a trip to Six Flags, so I had been there many times and really liked the area. The Raleigh-Durham region is great, but I wanted a big city that was close to my family, had a low cost of living, and had a strong technology presence. Atlanta fit the bill perfectly.

    Now came the next big iteration: We completely shelved our product that we had spent nearly three years building in order to build The Next Big Thing.  By mid-2002, we had accepted that we were never going to be successful selling installed PHP software to small businesses.  Instead,  a larger competitor with much greater resources would be selling a slightly modified version of our product to the same target audience.  It was a pretty easy decision to make.

    Our next move was to build a completely new, mid-market Java-based web content management system using the expertise we’d gained over the past few years. I promptly put an ad in Craigslist and hired the best Georgia Tech software engineer I could find. Being that it was late 2002, there were lots of good people looking for work. I found our lead developer within a month and we were off and running. Our new product combined XML, search engine optimization, and distributed server publishing in a formidable package. And because we were building a next-generation product from scratch, we were able to address all the mistakes we made in architecting the first product.

    Head on over to Iterate or Die – Part 3 to learn how we made this new product successful, and the sales and marketing iterations that ended up paying off.

  • Iterate or Die – Part 1

    I’ve been hearing a good bit of chatter lately among the startup community in Atlanta about how important it is to iterate in a startup. By “iterate” I mean the process of determining exactly which product, market, and business model will result in success – in short, finding out what works.  Of course, success means different things to different startups. Let’s dive into iterating in a startup.

    Lance Weatherby is a venture catalyst at ATDC’s Georgia Tech and a well-known blogger in the Atlanta tech scene.  He values velocity and versatility in team members, and on his blog Force of Good (http://blog.weatherby.net/) he coined the term velocitile to label such a person. I believe that you need both a good market and flexible people to be successful in a startup.  Atlanta is a phenomenal city for talented people that have these characteristics. With both of those in place, iterating is a natural and healthy part of building a company.

    My Atlanta-based Inc. 500 software company, Hannon Hill, makes mid-market web content management solutions for higher education and other industry verticals. It wasn’t always this way. I started the company in Durham, North Carolina, and my intial vision in December of 2000 was to provide a software-as-a-service (SaaS) application that would make it easy to update a generic, small business website, for $30 per month.

    The service worked with existing websites over FTP and provided a visual interface, similar to Windows Explorer, so that people could click on a file and edit it in a browser-based word processor. Upon saving the changes, the file would then be sent over FTP back to the web server, along with a backup version. The benefits of this model included:

    • No software to install on the client (web browser) or server
    • No initial up-front fee and a low monthly cost
    • Familiar file manager interface with word processor

    On top of these benefits, the software was as easy to use as web-based email. The only problem?  It was a complete failure.  I started out working on the company part-time and eventually went full-time within six months.  I learned several lessons shortly after going full-time:

    • The market wasn’t accepting of SaaS as businesses were used to purchasing installed software with a perpetual license
    • $30 per month per site did not provide a sufficient path with little capital to build a $1 million+ business
    • Customers needed significant hand-holding to get up and running even though the product was easy to use

    At that point, I knew something had to change. By August 2001 we had retooled the product to be an installed server application (it was always a PHP/MySQL app) and priced it at $1,000 for a 10-user license.

    Head on over to Iterate or Die – Part 2 to learn why we moved to Atlanta and if our iterating paid off.

    Note: This is republish of a previous series with some aspects reworked.

  • Post Mortem on a Failed Product

    Just over two years ago, at the beginning of 2008, we set out to build a web content management system with community functionality infused throughout — eCrowds. The idea was that companies would need a solution for facilitating product communities with the following functionality:

    • Content management
    • Forums
    • Blogs
    • Idea exchanges
    • Wikis

    We were solving the traditional challenges brought on by disparate silos of data with separate user authentication systems and inconsistent interfaces/template designs. 2008 was spent building the product and we launched it for our own internal customer success communities after nine months of development. It was a failure.

    Here are some of our lessons learned:

    • We spent way to much time building it for ourselves and not getting feedback from prospects — it’s easy to get tunnel vision. I’d recommend not going more than two or three months from the initial start to getting in the hands of prospects that are truly objective.
    • We made the product much too broad such that it did a little bit of everything. Unfortunately, it didn’t do anything exceptionally well. As an example, when we used it to replace our phpBB message board instance, our community was in an uproar because it lacked many of specialized features a mature forum offers (e.g. notifications and handling code examples in a comment).
    • The tools we were replacing were pretty cheap (e.g. each item was typically under $100/month), and we wanted to focus on the small-to-medium sized business, as that was our area of expertise, so we priced the product around $500 – $1,000/month. This caused another major disconnect — the integration costs were typically $10,000 to migrate an existing site into the system. With those integration costs, the sales cycle became prohibitively long/expensive relative to the lifetime value of the customer and the recurring revenue.
    • As the product became more and more complex, the performance degraded. In my mind, speed is a feature for all web apps so this was unacceptable, especially since it was used to run live, public websites. We spent hundreds of hours trying to speed of the app with little success. This taught me that we needed to having benchmarking tools incorporated into the development cycle from the beginning due to the nature of our product.

    So, to summarize, I recommend getting something simple into the hands of prospects quickly, picking out a narrow niche and doing it well before going broad, thinking through the economies of customer acquisition costs relative to the revenue generated, and making sure the app always runs fast.

    What else? What are some other lessons you’ve learned from a failed product?

  • The Pardot Name

    Way back in 2006 I had an idea for marketing software that would help with lead generation and management. Not owning any good domain names that met my criteria, I set out to find a domain name as the basis for the company name.

    After using several domain name finding tools (like BustAName.com) and not having any luck I went to Dictionary.com and typed in “marketing.” At the time, Dictionary.com returned translations of a word in 29 different languages. Nothing useful turned of for “marketing” so I tried the derivative of the word and did a search for “market.” Of the translations that came back, one caught my eye — pardot.

    Pardot is the Latvian (not Latin!) verb meaning “to market” or “to sell.” I liked that it was easy to pronounce and spell, so I headed over to BustAName to see if it was unregistered, and, much to my surprise, it was available. I registered it immediately for $8 and the rest is history.

  • Tools to Manage Tasks and Projects

    I’m always interested in learning how teams work within companies including tools they recommend. Methodologies like Getting Things Done (GTD) by David Allen has been great for me (I don’t follow it completely but did take several pieces from it). Today I had an entrepreneur shoot me an email and ask about project management software, so I wanted to share some tool ideas for both tasks and projects.

    Tasks and Projects:

    • Email – make a special folder to put action items in
    • Google Spreadsheets – this is my favorite when it is something simple and I don’t want much overhead
    • GQueues – good for task management with tight Google Calendar integration
    • Basecamp – great for lightweight project management that just works

    Of course, tasks and projects aren’t the same thing, so this list is just a starting point. My advice is to try out several tools and start using one to make your startup more productive.

    What else? What other tools do you recommend for tasks and projects?