Capacity Planning is all about trying to predict what the impact of future events will be on the current systems, and making sure that there are no bad outcomes.
Over the past few months, we have seen what may well become the most momentous decision for the UK take place… the Brexit Referendum. My work over the past year or so, has been to assess each of my client’s 200+ applications used for trading Equities and Derivatives and work out whether they could cope with the increased trading activities brought on by the referendum, or whether upgrades would be required.
But what exactly would be the increase?
This is often the challenge of Capacity Planning. How can you enumerate the impact of something that has never happened before?
Scenarios:
You could take the approach of modelling the impact of an increase of 50%, 100%, 150%, 200% of the current activity. This allows you to identify which of the applications would break “first”. You may find that there are a handful of applications that can only cope with an increase of 50% volumes, whereas the majority can cope with a 200% increase.
| Application | 50% increase | 100% increase | 150% increase | 200% increase |
| A | Y | Y | Y | Y |
| B | Y | Y | N | N |
| C | N | N | N | N |
| D | Y | Y | Y | N |
This gives you an immediate “hit-list” of applications to which you should pay attention for potential upgrades. But that is only if you think the impact of the future event (Brexit vote, in this case) will result in a increase of more than 50% on trading volumes.
BreakPoints:
An alternative approach, is to turn the whole modelling on its head. Rather than finding out whether a particular application will break when there is an increase in volumes of a fixed amount, one looks at the maximum capability of each application and report this in terms of the current workload.
| Application | Maximum Increase |
| A | 230% |
| B | 117% |
| C | 15% |
| D | 168% |
This view shows the weak-points in the environment, and is a more detailed assessment than that provided by the “scenario” approach above.
Yes, but, WHAT EXACTLY would the increase be?
Ah yes. This is still the $60m question. Obviously, no-one can know for certain what effect an unprecedented event will have… because it is unprecedented (the clue is in the name!). But we do have a couple of approaches open to us. We can find a “similar” event from the recent past and see what impact that had. In this case, the General Election of 2015. We can take that as a yard-stick and see what the delta in business volumes back then (ie: an increase of 150%) would be to the current business volumes (ie: still an increase of 150%, but on a higher baseline of values).
The other alternative, and always one to make the experienced Capacity Planner smile, is to actually ask the Business. After all, who knows the business of trading Equities and Derivatives better than the traders themselves? The reason this makes me (and others) smile, is that quite often the people closest to the sharp-end of the business, don’t actually know any more than you do about what the impact of a future event will be. Even when that “future event” is wholly within a business’s control, like a planned business strategy of selling twice as many widgets as they did in the previous year. Often the sales won’t come through… or the sales people will be super-successful and sell more than 2x. They aren’t going to STOP selling when they hit 2x (as per the strategic plan), they are going to keep on selling more and more!
So.. I hear you ask…. What approach did *I* take and did it work?
The joint decision was to go with a 100% uplift on previous peak volumes for all systems. I say JOINT decision, because while it was the proposal that I put to the business, I always like business assumptions to be “owned” by business. As the referendum results came through on that Friday morning, I was monitoring the systems throughout. The exit poll at 10pm told us that the result was very close, and that actually helped. For months the pollsters had been predicting a close result, so very few people had taken a completely IN or OUT position with their trades. There was no panic buying or selling as people looked to move out of their extreme positions, instead there was a steady trickle of trades as people consolidated what they had. The eventual uptick in trades was only around 30% at peak… something easily manageable.
So was this all a monumental waste of time? Is Capacity Planning just a futile activity with no beneficial impact to the business…. I would say NO… but to expand on that statement is the subject of a future blog.