David Cummings on Startups

Blog

Prioritizing Time for New Research and Development
Image by bill barber via Flickr

Two weeks ago I had separate conversations with different entrepreneurs about how they prioritize research and development in their companies. Each conversation was brought up independently but the respective entrepreneur lamented how as their companies had grown the speed with which new features were released slowed down — it was now hard to make time for R & D. Customers were requesting enhancements to existing features, the architecture team wanted to improve back-end items, and the support team wanted to make certain processes easier.

One entrepreneur was thinking of going to an 80/20 model where engineers spent 20% of their time on new features and 80% of their time on existing requests. The other entrepreneur was thinking of dividing the engineers up into different teams where one team would only be new features and the other team would working on existing functionality. There’s no easy answer.

Here are a few best practices when prioritizing time for new features:
- Continually re-iterate that every hour spent on a new feature is taking time away from some other change, and that that is OK
- Remind all stakeholders that you can’t make everyone happy but that by having a strong vision and opinion for the product you’ll minimize time spent on features that are too narrow in focus
- Work with sales, marketing, support, operations, and engineering along with soliciting input from customers and partners during the time allocation decision making process
Prioritizing time for new R & D relative to other requests is hard. The best thing to do is to make the necessary time, over communicate, and deliver great results.

What else? What other thoughts do you have on prioritizing time for new research and development?
May 14, 2011
Flashpoint – Great Y Combinator Clone at Georgia Tech

Image via Wikipedia

This week Georgia Tech announced their awesome new Y Combinator clone called Flashpoint. I first learned about the idea two months ago when I had lunch with Merrick Furst at my weekly spot (Houston’s across from Lenox Mall). Merrick described the next generation entrepreneurial environment as a Y Combinator-like program running right on Georgia Tech’s campus for people inside and outside the community. There’s no better education for an entrepreneurially-minded person to learn about building a startup than to actually roll up their sleeves and start building. Georgia Tech now has that.

Two and half years ago myself and a group of local entrepreneurs started a Y Combinator clone for Atlanta called Shotput Ventures. Our goal was and is to help improve Atlanta’s startup community through mentorship and seed funding. We’ve invested in nine startups already and have money to invest in several more. Go ahead, learn about applying to Shotput.

Flashpoint is great for the Atlanta startup community and Georgia Tech — I can’t wait to get involved as a mentor and I’m confident it will be successful. Sign up online to be notified when Flashpoint starts accepting applications.

May 13, 2011
How Fast is Too Fast to Push Out New Features

Image via Wikipedia

We like to push out new features fast. Very fast. On a typical day we might push out four updates of our software to production. Now, we don’t do continuous deployment but I’d like it if we did. Do we make mistakes? Yes. The great thing about our mistakes is that they are small, because we push the changes out in bite-sized chunks. With small mistakes come small issues that we can fix quickly. The more complex the change the more complex the fix.

We do a combination of code reviews, human testing, unit testing, functional testing, and continuous end user experience monitoring. This doesn’t catch everything but it catches the vast majority of issues and always ensures the core system works. We find that customers prefer fast product innovation with minor hiccups to much slower product innovation, and still have minor hiccups. To err is human, and we’re comfortable with that.

My recommendation is to move fast and build that into the core competency of the business. Startups win by staying closest to the customer and moving fast.

What else? What do you think about pushing out features fast?

May 12, 2011
A Strategy Session to Close the Deal
Image via Wikipedia

Earlier today I was talking to one of my star sales reps and he was telling me about a new tactic that he was using to help close business: a strategy session. The idea makes perfect sense. After engaging with a prospect, doing web demoes, and helping them feel comfortable with the process, you ask for them to purchase your goods or services, but they are still on the fence. In lieu of going straight ahead and asking for their business, certain prospects need a painted picture of success, which is where a strategy session comes in.

Here’s how the strategy session might work:
- Ask the prospect for one hour of their time to do a strategy session
- Provide three common best practice plans for your goods or services and have them choose one
- Take them through a series of questions around the chosen best practice they want to execute (e.g. increase PPC ROI through specialized landing pages with auto-responders)
- After you finish the phone call, send them a completed worksheet with all the information and execution plans
- Ask for their business and the opportunity to make them successful with the output of the strategy session
Strategy sessions are a great way to dig deeper during the consultative sales process, build trust, and ensure alignment of prospect goals.

What else? What are your thoughts on using a strategy session to close the deal?
May 11, 2011
A Killer Feature or a Killer Collection of Features
Image via Wikipedia

This one’s a tough one because you won’t know until you’re successful but the best products have a killer feature or a killer collection of features. The distinction here is terribly important as some products only need one killer feature (e.g. automatically identifying companies on your website) while other products need a collection of features that when combined make it a killer product (e.g. any one feature of Basecamp isn’t nearly as useful as the collection of features).

A killer collection of features is one of the main reasons we see feature creep in products as well as startups in stealth mode for an extended period of time — the entrepreneur/product manager believes it needs substantial functionality to be useful. These are the most common startups that die because they take so long to validate.

A killer collection of features is more difficult for several reasons:
- The functionality takes longer to build resulting in a greater chance of running out of money
- With more functionality comes more friction to adoption and understanding
- The chance of adding useless features grows while the chicken and egg problem of needing happy customers without a killer collection of features grows
My recommendation is to think hard about your killer feature or your killer collection of features necessary for success. The latter is much more difficult to achieve but is more commonly found.

What else? What other thoughts do you have on killer features vs killer collections of features?
May 10, 2011
Economics of a Roll-Up Strategy

Image by catchesthelight via Flickr

This morning I had the opportunity to talk with an entrepreneur that is starting the process of doing a roll-up for his market. After boot-strapping his company for the past 15 years he’s achieved a bit more than $10 million in annual revenues. Now, he’s made good money being the sole owner of the business and but he was ready for a new challenge: $50 million in annual revenue in five years through acquisitions and organic growth. His thinking is that his firm will be significantly more valuable to an acquirer with greater scale and more comprehensive offerings.

What are the economics of a roll-up strategy?

The current company, with $10 million in revenue, might have 10% margins (e.g. make $1 million/year profits), making it worth 4-5x profits (so, $4-5 million in value). Potential acquisitions are other firms in the space with lower revenues and are valued mostly based on profits, but also based on longevity of profits and growth rate. Thus, a firm with $2 million in revenue, 5% margins, and $100,000 in profits might be acquired for 10% of the equity of the combined entity even though revenue is 20% of the combined entity.

Why would the smaller company do this? As profits increase, company value as a multiple of profits increases due to the potential for greater economies of scale and sophistication of a potential acquirer such that a company with $5 million in profits might be worth 7-8x profits ($35-$40 million in value). So, for the smaller company, their same $100,000 in profits might be worth double (e.g. going from a 4x profit multiple to an 8x) due to being part of a large company. This happens all the time.

Roll-ups are extremely difficult and the best acquisitions happen with aligned corporate cultures. The economics make sense for entrepreneurs that are ready to hitch their wagon to someone else’s train and believe that the opportunity for success and the weighted expected outcome is higher.

What else? What do you think of the economics of a roll-up strategy?

May 9, 2011
Pods in R & D at salesforce.com
Image via Wikipedia

A friend of mine from college joined salesforce.com (yes, all lowercase is the proper spelling — it’s a late 90s dot com thing) right after he finished business school and has been there for several years. At last year’s Dreamforce conference I was asking him about their engineering department and how they structure on-shore/off-shore software development. He said all the R&D is done in San Francisco in a pod-like setup similar to yesterday’s post on pods at Rackspace.

Here’s some info on pods in R&D at salesforce.com:
- There are roughly 28 different teams/pods
- Each team focuses on a specific module or piece of a module in the product
- The pods have one product manager, roughly five engineers, and a QA person
- Each pod does a daily stand-up scrum meeting led by the product manager
- The large number of product managers allows each of the small teams to stay close to the customer and minimize the telephone game where details get lost in layers of bureaucracy
The pod approach makes it easy for salesforce.com to scale their R&D by adding more and more pods as the business grows while staying agile and innovating quickly.

What else? What do you think of the pod approach to scaling R&D?
May 8, 2011
Pod/Matrix Team Approach in Startups
Image via CrunchBase

A couple years ago the co-founder of Rackspace told his success story to the EO Atlanta chapter. One of the takeaways I enjoyed from the event was the concept of the pod/matrix approach. Morris, the co-founder, recounted Rackspace hitting several hundred employees and feeling the growing pains trying to service customers with fanatical support. They had dozens of people of in their support department, dozens of account managers interfacing with customers, and dozens of people in accounting to field all the billing questions.

Departments where having difficulties with more and more layers of management and customer complaints were increasing. In addition, the executive team was really concerned as to how they would scale the entire business from a few hundred people to a few thousand. The solution was a pod system where teams of five were built to handle the majority of customer questions quickly and efficiently. Pods were built to match the most common customer requests:
- Two Tier 1 support reps
- One Tier 2 support rep for harder technical issues
- One account manager for sales
- One accounting clerk for billing
With these pods in place, scaling the business was simply a matter of having more and more of these teams as they added more and more accounts. Problem solved.

What else? What are you thoughts on a pod/matrix system like this?
May 6, 2011
On the Startup Path

Image via Wikipedia

As I was rushing to the airport tonight to fly from Chicago to Atlanta (I’m prone to be late) I had time to reflect on the train. You see, sitting on the train to Midway Airport there was nothing I could do to speed things up to catch my 6:45 flight. I had my goal in mind: make my flight so as to get home at a reasonable hour for a weekend with my family.

Being the startup geek that I am, I equated the experience with entrepreneurship.

Entrepreneurs have a goal of building a successful company. Once the goal is crystallized, much like my more simple quest to get home to my family, you start working backwards and planning the steps to achieve it. During many parts of the entrepreneurial journey there are situations where you put yourself in a position that you are at the mercy of others around you, no matter how much you’d like to speed things up, similar to how I was sitting on a train hoping to get to the airport faster.

The most important takeaway is there are many things you can control and many things you can’t control. Setting your sights on a target, controlling what you can control (that was a favorite line of Andre Agassi’s coach), and putting yourself in a position to succeed makes all the difference.

What else? What are some other parts of the startup path?

May 6, 2011
Why the Proliferation of Successful Email Marketing Companies
Image via Wikipedia

Yesterday I highlighted almost a dozen email marketing companies with at least 100 employees or $15 million in revenue. Yes, that’s right, a dozen highly successful email marketing companies in the United States alone. That’s a ton of companies. The big question is: why hasn’t the email marketing industry gone the way most technology markets with a winner take most outcome in a few industry segments?

Here are a few ideas why there’s a proliferation of successful email marketing companies:
- Email marketing, unlike most markets, can show a return on investment within an hour of using the tool
- Most companies and organizations need email marketing leading to an abundance of specialization by vendors and a large overall market
- Recurring revenue businesses with high gross margins, like email marketing, make it easier to grow and build a sustainable business once you’ve crossed the desert (say $2 million in revenue, which is incredibly difficult to achieve)
- A lack of network effects and the multitude of spam challenges and email clients make for fewer reasons to switch providers as an end user, assuming things are working well
Email marketing is an unusual market for these reasons and more. The proliferation of successful email marketing vendors shows no sign of slowing down.
What else? What are some other reasons there are so many successful email marketing vendors?
May 5, 2011