Ben Allfree :: Painless Programming

Several clients are interested in a Twitter clone. Part of the service I provide and the experience I bring is risk assessment. In this article, I talk a little about how to assess the business risks involved in your next project idea and how I can help.

A Twitter clone is a timely example. Each client has his own twist on Twitter. Maybe it’s a specific chat vertical. Maybe it’s not chat, but some other kind of social communication. But basically it’s Twitter in the sense that it’s real-time multi-protocol instant messaging that gets echoed to a web site.

Before Twitter came along, I didn’t even understand that people wanted this. I still don’t personally have any use for social networking applications. At least not the way they are today. You can find me on mySpace, Facebook, Twitter, LinkedIn, and more. But I never use them.

It helps if you can be the target market for your idea. If you can be your own best customer, then you have an intrinsic understanding of their desires. But I suspect that is a rare luxury. After all, we are producers not a consumers. As a rule, producers do not think like consumers.

As a producer, I wonder to myself: Twitter already did Twitter. What is the value of adding a little spice to it? Would the time be better invested in a new idea that comes out of left field like Twitter did? What are the risks involved in making another Twitter?

I can generalize project risks into three categories. Part of the value I bring to engagements is teasing apart the nuances of the business case. I need to see a good business case before I will take a project.

Why Projects Fail

90% of projects fail to see the light of day because of a false premise early in the life cycle. I like the other 10% and want your project to be there.

Fred Brooks (Mythical Man Month) famously wrote that more projects fail due to a lack of calendar days than all other reasons combined. Projects simply run out of time. If you can do everything right, you might be lucky enough to run out of time on your project. However, I think there are some good predictors of success (or failure) besides waiting for the calendar to end.

You are probably imagining a false premise is something like an improper data field length or type. Oh, no. The kinds of false premises that kill projects are much more obvious.

1. Missed risks

If we build it, they will come.

This is probably the most common one. I am guilty too. You think long and hard about all the nuances of your idea. You figure out all the angles. It’s a perfect idea. It can’t miss. All you need to do is build it.

So you build it. They don’t come. Repeat and rinse. What went wrong?

My advice for all entrepreneurs is to find three customers. Not prospects; not people who say they will buy it. You need to find three customers who will put money down for the solution. Failing that, find three customers who are willing to donate staff time toward a pilot program. It needs to be enough staff time that it is a real concession on their part. If you can not find three customers, I like to say that the business idea can’t find traction. It may be time to move on to the next idea.

I use the magic number three because it seems to be an effective filter. One is definitely not enough. Two could just mean one has a friend. But three…three seems to adjust for any bias and makes for a good sample.

2. Inaccurate risk assessment

Risk is pervasive. You can avoid risk altogether and in that case will not need my services. But if you want to make something, you must manage risk.

You can take steps to contain risk. The first step to risk containment is risk assessment. You need to know the probabilities involved. You have to guess, and it’s scary. You have to invest based on your expectations of the future. A faulty risk assessment occurs when your prediction model does not properly mirror the probabilities of the various outcomes. Key point here: you predicted the outcomes, just not the likelihoods.

Here is a very common example. Client A wants free customizations as part of the deal. He says he will walk if you don’t give in. So, you rationalize that there is a “good chance” - let’s say 90% - that future clients B and C will want the same feature. You tell yourself the smart move is to bet that you can sell clients B and C more easily because of the investment you made for client A.

Suppose that the 90% chance you imagined actually turned out to be a 10% chance. In other words, you had to close 9 more deals before you found 1 that actually cared about the feature you thought everyone wanted. Making decisions as if 10% were 90% is a good way to kill a project.

Having someone around who has seen many projects can be invaluable. It’s your own form of an advisory board. I see what works, I see what doesn’t, and I help you accurately contain risk.

3. Ignorning estimates

The first way to kill a project is to miss elements of risk. The second way to kill a project is to inaccurately assess riskiness. The third way to kill a project is to ignore risk.

Suppose you need a new water pump for your car. You call your regular mechanic, who you know and trust, and he says $800. You get another estimate of $700 from someone else. Finally after much searching, you find someone who will do it for $450. All things being equal, you should go with the $450. But you know things are not equal. What are the risks? You probably won’t know. You can be assured through experience and past transactions that your regular mechanic has a 99% chance of getting the job done right. The rest has to be estimated based on very little information. Even if you learn about the risks (different brands of parts, different installation skill level), can you accurately assess them? And if you do assess them and find that the $450 deal is fairly risky, will you be too tempted to ignore it?

The water pump example is contrived, but here is the story of a $250,000 mistake which is true. A client called and asked for an estimate. I estimated the project size to be roughly $1 million. He thought it sounded high, so it sought an alternate estimate and found a reasonable discount at $850,000. He continued his search until he found someone who gave him the answer he wanted: $250,000. What has this client gained? Has he really found someone better, or has he simply found a smoother talker?

The client made the obvious choice: he chose the low bidder, which was still a lot of money. After all, every vendor appeared to have similar credentials, similar teams, the same information and could seemingly render the same quality of service.

The project crashed about 1/4 the way through, quite predictably to the rest of us. Who is to blame? The vendor? Perhaps. But the client chose to ignore information that didn’t fit what he wanted the situation to be.

If you have fewer than 15 years experience running projects, you have to choose a provider you can trust with your livelihood. Custom software is too complex to run with anything less.

This brings me back to Twitter. How can we make a Twitter clone while avoiding these three major causes of failure? The answer isn’t in this post. Instead, my assertion is that I can help you navigate the causes of failure, contain your risk, and if a Twitter clone business idea is sound, take your idea to market better than anyone else.