Saturday, December 03, 2016

Building a business from a great idea, some future Monday

[Rough draft - the more upto date version is on medium]

It's Monday, it's the year 2025 and I've woken up with a great new idea. [Actually, it's a pretty lousy idea but hey, I just spent five minutes on the scenario so let's just assume it's great.]

I'm going to create a recommendation engine for stock picking based upon the mood of the internet. I quickly scribble down the user needs "make profitable trades" and "know what to buy" and write a basic map whilst grabbing breakfast. I have my map, I know the basic components that I need - the recommendation engine, trade feed etc. I start work at 8.30 am.

I know Amazon provides one of the components as a service (the lambda platform) and several others can be found as AWS lambda services in the marketplace. The company I work for also provides a stock portfolio service. I mark up what I think we can use and what we need to build - the recommendation engine and the mood system.

It's 9.20 am, I send the map of to our spend control group. They act like an intelligence gathering organisation, collecting all the maps of everyone, comparing them and giving some feedback. They build profile diagram by finding common elements between the maps.

I get a reply by 10 am with some details. They send me the profile diagram below. It seems some other team in the company has built a recommendation product. I'm the only person thinking about a mood system. In general, I'm roughly where everyone else, However, 16 different teams are using trade feeds and everyone else is using some well developed lambda service and apparently everyone else is using some utility service for a trading engine. I check click the details, it's Amazon.

They've also sent my map back, slightly modified. Ok, well at least this is not the like the bad old days of 2015 where my company had thousands of duplicated systems and endless rebuilding of stuff that already existed.

It's 10.15 am. I start thinking about some metrics. The trade feed system is going to be providing trades to the recommendation system. Each one will need a call to the mood system, the risk system and so on. I start marking out where all the metrics are.

It's 10.45am. I flip a switch and the map is converted into a business model. The same components, the same links, the same metrics. I start running a few scenarios, checking that I've made a truly variable cost business model. It's 11.30 am, I send the map and the model to finance. They come back with some helpful comments [in this case, it would be ... and how do we make money? but then again the scenario took me five minutes].

It's 12.00 am. I send the maps and the improved model to the executive group. 15 minutes later I get the go ahead and a small budget of $20k.  I know from the spend control profile that some other cells are already building this stuff. I give them a call, tell them what I'm upto.  They already know, spend control told them.

I know from the spend control profile that there is a group building a recommendation engine. I send them the map and model and outline my idea of adding a mood system to recommendation. We have a quick call and they're up for it. We agree a metric of value for charging - everyone uses worth based development these days. Most of the stuff is already built and provided as services. I just need a cell of pioneers to build the mood system, whatever that will be.

I update my map with the organisation structure and load it with the build map and financial model to the company's job portal. I wait. 

Our company operates in cells, using pioneers, settlers and town planners. We live in a constantly changing environment. Watch the movie! I love it.

Because of this, we always have pools of people training and looking for their next cell to join. It's 1pm and no-one has responded. I'm getting worried. I'm looking at the other exciting projects on the jobs board. Out of the blue, by 1.30 pm I've nabbed two pioneers willing to give this a go. They sign up.

We're off to the races! Of course, HR is constantly monitoring the flow of components through the maps, the cells being formed, whether we need more pioneers, settlers and town planners. This goes on in the background. They're checking what we build vs what we buy and whether we have the balance of attitudes. Long gone are those old days where dullards would try to convince us that a company could have one culture. Long gone are those days we were weren't looking for the right skills (aptitudes) and the right attitudes. HR is on a bit of recruitment drive at the moment, we've been lacking enough settlers especially in finance and engineering.

We start cracking away with the project. We build the mood system, add it to the recommendation engine and start watching whether consumer use it. We start monitoring flow in the system, where's the money going, are there bottlenecks, how are we doing on those metrics?

Of course, we're not the only ones monitoring. Part of the spend control group looks after strategy and they are already looking at the maps for new opportunities. One of the things about our stock portfolio system and recommendation engine is other companies build on top of it. They can measure consumption of the service to identify future trends. But they're also watching how the mood system is going, maybe we should provide it as a service to others?

They notice the mood system is picking up. They decide we should push its evolution towards more of a utility. It'll need development and in this case, they decide an open approach is worthwhile. We've only got going with our system and I've noticed a new project on the company job board to turn the mood system into an open sourced project. 

It's 5pm. I'm in a good mood. The mood cell is up and running, it's even growing with an open source effort. The changes to the recommendation engine are working. I have a relaxing evening, get a good nights sleep.

It's Tuesday, it's the year 2025 and I've woken up with a great new idea.


The scenario was put together very quickly and is only an illustration designed to explain one thing. If you use a map then there is no reason why operations, build, strategy, finance, HR and other groups can't happily work together without miscommunication, misalignment, duplication and bias. All of the above diagrams I've used in one form or another across multiple groups in a business over the last decade.

There currently is no integrated tool for doing this but I strongly suspect that our future development, operation, HR and financial tools will be combined together as above through the use of mapping of some form.