The more members you add to a team the harder communication becomes. To that end, it's sometimes hard for me to understand why startups are so focused on growing their teams out during the early stages.
For all the talk about how good communication is key for a team's success, I have a counter idea: Avoid having to communicate in the first place. Communication is not the ideal to a team's success; less team members to communicate with is.
The closer a team size moves to one, the more efficient its productivity.
Jeff Bezos likes to refer to the ideal team size as "two-pizza teams:" any team that is small enough that they can be fed by a couple of pizza pies, is a model of efficiency and accomplishment. Anything larger is not.
Image courtesy of Randy son of Robert
The medium is the mess
Communication is actually bad. It inherently involves a loss of information. The more communication that is needed, the less of the product plan will be efficiently implemented according to the original vision.
Consider the impossibility of trying to tell a friend about a wierd dream you had the previous night. Can you convey every single detail of the dream before it fades? Of course not. You have constraints (like time and memory), so you cut out anything which is "insignificant."
The same holds true for any project plan. The larger the amount of people to convey information to, the less efficient you can be at it. Explaining to one person every detail of the plan is tedious enough. Imagine having to do the same for multiple people. Daisy chaining the information so that it is passed on from executives to managers to smaller groups has its own problems: Information loss and corruption at multiple points. It's the age-old game of "operator," only with results that are not nearly so funny.
When there is one person both running and operating the entire show, you have 0% communication efficiency loss. The vision is designed and implemented exactly as it was originally conceived. Add a second teammate and you automatically introduce inefficiency into the equation. With each new person added to a team, the potential for communication efficiency loss gets worse as each person creates failure points with every other person. Once you start getting beyond 8 team members, the efficiency loss becomes so great that it can only be made up by throwing additional resources at the problem. In other words, you are not going to see double the output from a team of 15 people as you will with a team of 8 (even though you'd expect it on paper). In fact, you'd be lucky to see even a 25% increase in output, even though your team size has doubled.
Keeping your team small
So what's the overarching lesson? You don't need a huge team to successfully launch a start up. In fact, your chances of succeeding are better, the smaller your team size. You cut out as many communication points of failure as possible and keep your startup costs down.
So how do you keep your team small?
* Choose a project that is simple to implement. Don't try to create a complex suite of applications. (Yeah, I'm a hypocrite). Focus on solving a single problem. Philip Kaplan made email more efficient to use by stalling it instead of managing it. Dead simple approach and a great idea.Take the easier approach when possible.
* Choose people that can wear multiple hats. Can your designer code? Can your programmer manage a community? Can your marketing guru fund raise? Can one guy do it all?
Image courtesy of Mike Burns
* Document everything. It's obvious that you will need a business plan. What's not so obvious is that you should also document the seemingly mundane; methods used for team communication, methods used for integrating with potential partners, methods used for keeping a company blog up-to-date and interesting. All documentation should be available via a central location. A wiki can work really well for this purpose. Good documentation lessens the loss from communication failures.
* Arrange your workspace in common areas. Segregating your team in different offices is a recipe for lost communication data and with it, a need for additional people. You'd be surprised at how many roles can be shared by multiple people, so long as they have the ability to communicate instantly and unimpeded with each other. Put people between walls, and those shared tasks will need to be managed by additional team members.
Examples
The following are examples of two-pizza teams that generated some of the most popular community content sites online:
* Fark
* Worth1000
* Newgrounds
* SomethingAwful
* Delicious
* Metafilter
* Etsy
* Reddit
* Flickr
Know any others?
I've been watching the steady decline of journalism since the Internet began replacing print and television as the main provider of news, with a seething disgust.
Today's CNN top story put me over the edge.
Here's the title and article summary for those of you without images in their feed readers:
6-year-olds forced into sex for food, group finds
A poor Haitian girl could get $2.80 and some chocolate, she told a European charity. All she had to do was perform a sex act on a humanitarian worker. She refused. Her impoverished friends did not. Her story is one of many in a report titled "No One To Turn To" -- which chronicles allegations of charity and U.N. workers abusing children.
But if you read the actual article you see not a story about a 6-year-old being raped (it's a mere footnote in the article), but the following:
In the report, "No One To Turn To" a 15-year-old girl from Haiti told researchers: "My friends and I were walking by the National Palace one evening when we encountered a couple of humanitarian men. The men called us over and showed us their penises.
"They offered us 100 Haitian gourdes ($2.80) and some chocolate if we would suck them. I said, 'No,' but some of the girls did it and got the money."
This bait-and-switch is so misrepresentative as to be grotesque. The story changes from "Humanitarian workers pay teenagers for sex" to the more sensationalist "6-year-olds forced into sex for food."
The biggest problem with news being disseminated online is that there is no geographic isolation (as is the case with both print and tv), which means that every local news network is in competition with every other news site on the planet. Ratings are driven by attracting as large an audience as possible... and most people care more about Angelina Jolie and Brad Pitt's married life than how many people vanished from a Darfur town this week (hint: all 30,000).
Blue = Angelina Jolie, Red = Darfur.
Newspapers can't help but take notice.
Accurate facts used to be the hallmark of professionalism in newspapers. Editors would strive for it. Now even the most respected newspapers on the planet are feeling the creep of ratings greed and with it, an end to an era of accurate, informative news.
I'm not the first to make this complaint and I have no connection to the journalism world except for some past memories of running my university newspaper. More important people than I have used larger podiums to disseminate the same message (and gotten flak for it).
I understand newspapers are a business to run and profits are driven by advertising. I also understand that newspapers are the fourth estate, keeping world governments in check through the power of disseminating information. With every sensationalist article they run, every inaccurate headline, every news story that breaks the papers' traditional format because of a previous story's high Digg count: they are relinquishing that power in the name of profit. There has to be a balance.
As maddening as watching reputable brands peddle sensationalism might be, I actually have a bigger worry: Newspapers are clearly noticing how much Digg traffic certain articles are receiving; a fact that is certain to play a role in influencing the editorial direction of future stories (or at least their headlines).
I predict a future in which USAToday announces America's next president with the formulaic made-for-Digg headline, "The #1 Most Elected President Ever, in the 2008 Election."
The following are some random tips that I've learned through my experience founding and working at several startups. Take the opportunity to learn from other people's mistakes -- the best advice is that which is learned on someone else's time!
I. Relax
Image courtesy of audi_insperation
Work time should end at a certain point during the day. Period. Just because you are focused and energized by a specific task doesn't mean that you should continue to work on it until you conk out on your keyboard at 4 a.m.. The rest of your productive week will be completely ruined. Instead, stopping work on a project you are really into will give you a jump start on your work the following morning, and your energy will hopefully last through the day. Finally, turn off your cellphone / blackberry and switch gears: kick back to spend some time with family, read a book or watch (something intelligent on) TV.
Now, obviously, this doesn't apply to deadline days. If you have a project due the following day, you should never blow it. But, as general rules to live by, budget your work and plan ahead so you don't work into the late hours of the night, and always set aside time for non-work related activities.
Your work hours will be more productive if you are well-rested and well-rounded.
[Note: If my wife sees this tip, she will, of course smirk, because this is the one tip I break repeatedly, every single night (including right now).]
The full list after the jump...
Continue reading full post ...
Shuttling between tech meetups, VC meetings and conferences, I realized that there are so many types of entrepreneurs that it would be a mistake to group them all under one general umbrella. So I tried to correct that horrible wrong with this list of all entrepreneur types from A-Z.
Classic example of a fantrepreneur photo courtesy of Phil Hawksworth
ALL ENTREPRENEURS FROM A - Z
Againtrepreneur
Just sold their 5th company in 3 years.
Bumtrepreneur
They litter the streets of San Francisco, sleeping in doorways and begging for spare change. They individually make more money in a month than most Web2.0 companies.
Can'trepreneur
5 failed startups and it's probably worth revisiting that 9-5 deskjob.
Don-trepreneur
The Godfather of investors for Entrepreneurs. You probably want to check your termsheets carefully for the clause on broken kneecaps.
Entrepreneur
General class that accurately describes only 5% of the groups on this page.
Fantrepreneur
HOLY CRAP KEVIN ROSE JUST WALKED INTO THE ROOM HAI KEVIN I LOVE DIGG CAN YOU GIVE ME TIPZ FOR MY STARTUP PLEAZE?
G-Z available after the jump...
Continue reading full post ...
There is so much to learn from the implementations of design concepts in games that can be applied to non-gaming.
Almost everyone working on Aviary is an obsessive gamer. Besides for offering us a convenient (and violent) way to deflate, we also find a lot of inspiration in the underlying design. That's how I justify it to my wife, anyway.
Valve Software
Two games we are currently obsessed with are Team Fortress 2 and Portal, both created by Valve Software (most famous for its Half-Life series).
Valve's success as a software company stems from their philosophy of not compromising their story narrative for the sake of interactive gaming elements. Ironically enough, by forcing their game developers to work within more difficult parameters, they end up building better interactive elements as well!
Robin Walker, Valve employee and creator of Team Fortress 2 puts it best in the in-game commentary:
Holding ourselves to strong design principals can often force us to come up with better solutions than taking the easy route.
Lesson Learned: Limitations generate creative solutions.
Team Fortress 2
Team Fortress 2 is a game where users can choose between becoming one of nine different characters, each with unique abilities and limitations. Players will adopt different characters so that their team will be balanced properly. Having 9 different types of players running around on a field is plenty confusing. Teammates would have a hard time identifying and working with each other and finding certain characters they need (for example, a medic to recharge their health). That's not a problem in TF2 though, because Valve took the novel approach of designing the characters to be physical caricatures of their abilities, instantly recognizable by their silhouettes. Confusion is completely minimized.
Andrea Wicklund, another Valve employee, says:
The more your art direction can use well-understood visual representations the less work you have to do to explain you game elements.
Lesson Learned: Good design lies in the shapes.
How to apply it: Make sure your applications interactive elements (i.e. button icons) are all identifiable by their shapes alone. Exaggerated shapes are easier for people to identify and understand. Here's a great reference point.
Portal
Portal is a first-person shooter game where users are given an obstacle course and a single weapon: a gun that shoots portals. A user can open two portals simultaneously, and walking through one makes the user exit the other. The brilliance of Portal is in understanding that physics continues to operate normally in the background and must be used in helping navigate the obstacle course. For example, shoot one portal in the ceiling above you and one portal in the floor below you and you will begin to fall straight down between the portals (ad infinitum), increasing speed as you hurtle towards terminal velocity... pretty useful if you are trying to generate speed to hurtle yourself to a previously unreachable platform! 
What's amazing about Portal is that there are no true enemies or weapons in this first person shooter. It's nothing more than a mental challenge that defies you to solve puzzles by throwing away everything that feels right to you about physical interaction with the world. It is a completely new form of interaction with a previously existing genre of gaming.
Lesson Learned: Innovation can be found in minor refactoring.
In layman's terms: You don't need to reinvent the wheel to produce something completely novel.
There is a prevailing attitude amongst creators that it is better to release a prototype and rebuild it the right way if it becomes popular then building it the right way from scratch.
That's foolish, but it's also largely unavoidable, since you can never really know what problems exist with your prototype until their has been mass adoption of it. The Catch 22 is that if your prototype does get properly adopted, it's already too late to rebuild it from scratch.
There's a line at your door and they want immediate attention.
People don't wait around for you to change. Copycats spring up. Investors see revenue potential, customers attention spans are short. Your direction becomes one of patching the prototype and playing catch up with your scale requirements. Redesigning is impossible.
Initial design flaws seem obvious to us in hindsight, but they rarely are at the beginning. People don't think of a prototype as a prototype until they have to patch a new bug.
Take the Y2K bug: Why wouldn't computer programmers have thought about dates after the year 2000 when designing the first computer languages? It wasn't even that far off!
The truth is that most creators simply don't expect wide adoption and it is a lot easier to build a product without worrying about scalability. That's unfortunate, but it doesn't just apply to programmers writing applications.
There are so many real world examples of hacking being the only way to address a design flaw because mass adoption prevents re-addressing the underlying issue.
New Orleans
Why would original developers choose to build a city on lower ground than sea-level in an area prone to hurricanes? Why not import more land to raise the city above sea level?
The answer is that they probably didn't intend to build a city as big as it became and levies seemed a more realistic, economical solution in dealing with a smaller city. They never considered problems of scaling, because it's impossible to predict population growth.
A city grows around a current need (i.e. access to maritime trade) and remains standing once that need fades away. Then it is up to the new residents to continue developing on an infrastructure that was never intended to support continuous growth.
The design flaw becomes an inherent, unchangeable limitation based on mass adoption. You can't rebuild or move the city itself once it becomes obvious that it simply can't scale.
So New Orleans planners hacked their inherent design flaw by building a levy system instead of raising the city up higher in the first place. Woops.
English on the Internet
English has become the global language of the Internet.
Why? Because English-speaking people invented the prototype and didn't consider the global potential for it.
Because of this lack of foresight we are stuck with browsers that do not readily accept foreign language characters for URLs. Imagine how that limits countries where English isn't a first language. Take a look at the URLs for Wikipedia articles in non-English languages. Talk about non-intuitive usability!
And how wonderful for Google and other search engines that are the Hack to this terrible design flaw.
QWERTY Keyboards
The common layout of keyboards that almost all computers come with, known as QWERTY, causes problems of inefficiency and fatigue as people type. A more ideal layout is known as the Dvorak layout. It places keys in positions to improve efficiency in typing to almost double the current speed, but it's hardly been adopted at all.
So why did we even use a QWERTY layout in the first place? Because the concept of typing originated on a now extinct need (manual typewriters), and mass adoption of that character set has persisted a limitation.
The QWERTY layout was designed so that successive keystrokes would slow down typing and alternate between sides of the keyboard so as to avoid jams in typewriters. 
And now we're stuck with a mass adoption of a way of doing things that can't be redesigned. The Dvorak keyboard is a redesign instead of a hack and therefore it will never be adopted by the masses.
Others
Can you point out other real world design flaws where mass adoption limits us to hacking instead of redesigning?
I'd love to hear them.
One of the first steps in planning out development of your product is establishing your target audience.
In every industry there is a steep slope that represents market share and an important strategy decision has to be at which point on the slope do you enter? 
Target too high and you're catering an to important niche user base, but won't hit the broader consumer base for a while. Too low and gravity will keep you from ever making it higher up the slope.
In picking where to enter the market, most businesses base their decision on immediate return. It comes down to which portion of the market will give them the largest base for the lowest cost. Therefore most companies will take the bottom-up approach, targeting the bunny slopes first with a product that has broad consumer reach and lower costs to develop, before moving on to (or possibly choosing to pass up on) a more targeted and expensive market.
I think that can be short-sighted.
With Aviary we are taking the more unconventional top-down approach: Targeting a niche of semi-professionals with our tools first and then streamlining versions of our tools down for the masses, once advanced users are happy with them.
Why? It boils down to long-term branding effects and better software design. When your brand takes on elite connotations because it caters to the elite it becomes desirable to the masses. Sure, out of the gate you become the underdog as far as overall market share goes, but as time goes on and you begin to diversify you are left with an extremely strong brand, one that can be easily adapted for markets with broader, less-targeted ranges because of the branding strength and inherent software power. It's a matter of removing and simplifying features, not hacking onto a design that is not intended to be scalable.
The added benefit to a top-down approach is that you have nowhere to go revenue-wise but up, since your market base gets larger as you go further down the slope. Your market share only broadens as your products target range does.
The flip side is that companies that took the bottoms-up approach to grab as much overall market-share as possible often have nowhere to go but down. Professionals are bored by bunny slopes.
Case in point, recent news that Apple is finally worth more than IBM.
It's pretty tiring running a small web business - you're forced to constantly change your identity and mode of thinking.
- Mocking up a presentation? Enter Designer mode.
- Crunching figures on a spreadsheet? Enter Accountant mode.
- Tweaking your code base? Enter Programmer mode.
It's not the physical adoption of an identity that is exhausting. When I'm in a certain role, I am in a zone, focused on my specific task and nothing can distract me. But ask me to switch identities and my brain goes into shut down mode and I want nothing more than to procrastinate, anything but to don a new identity. The act of switching identities is simply exhausting.
I imagine that it's very much the dilemma Bruce Wayne faces every time he changes into Batman.
Actually being just Batman? Or even billionaire playboy Bruce Wayne? That's simply kick-ass. It's the changing process that's time consuming.
Think of the amount of work involved in slipping into your private study, finding the right book to trigger the secret entrance to the Bat Cave, removing your tuxedo, donning 100 pounds of protective gear and armor, ripping nylon tights on over your hairy legs, remembering to stop hitting on the ladies, remembering to start hitting on Robin...
It's not like Bruce Wayne can just slide down a pole and instantly turn into Batman, right?
On second thought... strike that.
So how does a small business operator cope? I think the best thing you can do is to force yourself into a majority role-a-day mode. Don't try to change your identity too many times in one day, unless you absolutely have to. If you designate specific days for specific tasks instead of chunking your day into smaller pieces spent on multi-tasking, you begin to spend more time in the zone and less in mental transitioning.
Friday for me is the day I pay my bills and do accounting; Weekends are for thought process and planning; Monday's are for networking follow-up; Tuesdays and Thursdays are wild cards, usually used for programming, UI testing or design.
And Wednesdays? On Wednesdays it's business time.
< view newer posts | view older posts >
Try Out Aviary»
Aviary is a powerful suite of browser-based design tools for people who create. Head on over to the Aviary homepage to try the applications out and learn more. It's free to try our tools or sign up!
RSS
Email