So you want to be an ITSM consultant?
Why not pit your wits against Peter and analyse this financial services case study. Do you agree with Peter’s analysis? How would your diagnosis differ?
Case Study Introduction – The scenario
By Peter Brooks.
There’s an interesting article on Financial Trading systems here. It raises many questions, and I thought it would be interesting to see how a consultant would tackle it if asked by an organisation to give service management consulting on the future of its trading systems.
It’s often easier to tease out the important matters by looking at a specific example.
So, here’s your mission: You’re new to the world of consultancy though you’re fairly familiar with service management and you have a background in IT. A friend of yours in the financial world has recommended you. He has convinced the trading manager that you can help him improve his bottom line with consultancy exercise. What would you do?
Scope – terrifying!
You’re completely new to this world. You’ve seen both the ‘Wall Street’ films, and you’ve got a picture. People in this world are seen as brilliantly clever, hugely impatient, highly aggressive and not the sort who lose many night’s sleep worrying about the ethics of their actions. So that’s the stereotype.
So the first question, as an inexperienced consultant is; ‘Should I take this job at all?’. A good question. You should ask this about every consultancy job you ever have, actually, and consider the answer seriously. Not just because it’s the right thing to do, but because it will help you think out your approach before the engagement and help you understand the big picture of what you’re going to be doing before you are too lost in the fast moving detail to have time to spare for big picture thinking.
A list of pros and cons helps – particularly if some allow you to rule it out or rule it in depending on what you hear when you first meet your client.
- If it works out it’ll be great for the CV, give you a useful insight into the financial world and might even make a difference to the world as a whole, in a positive sense.
- They know that you’re young and inexperienced, so they’ll probably ignore what you say anyway – it’s more being done as a favour, so, if you get it wrong, you may not do too much damage.
- It’ll impress everybody to know that you’re working in such a high-pressure environment, particularly if they don’t know all the details.
- If somebody does take you seriously, and you get it all wrong, then you might contribute to yet another crash and huge misery for millions, if only in a small way.
- Some people may be impressed, but some of your more environmentally friendly friends may never talk to you again.
- If your scope is limited to advising on one particular financial system, financial service or IT system, you truly don’t have the knowledge or experience, so you must say “no”.
- If you do discover major risks, serious ethical problems or other dangers, you must make sure that you are not made complicit. So you can ask your client, at the first meeting, what you should do if you do discover these. If you are told that you’d have to cover them up, then walk away. If your answer is that you should produce a private report for the board if that happens, or similar, then things look set fair.
- Will you have enough time to find out what you need to know? Make sure, at the first meeting, that you’ll be able to meet and interview all the major stakeholders to get a proper understanding of the situation. If you’re expected to only talk to one or two people, then you might accept the job, but make sure that your result must be accepted as tentative and a first-pass effort.
The last three are the important ‘cons’. At least now you have something to talk about at the first meeting.
Expect to be asked how you will approach the assignment at the first meeting. This is easier than you think because you’re not going to ‘wing it’ on your own, you’re going to be using industry standard advice on service management to guide you.
First you’ll need to understand what the business is actually doing and why and what your role is going to be in helping them. This is known as ‘enterprise analysis’ and, in this case study, we’ll be using the article from The Register.
Next you’ll need to identify the stakeholders and understand their various requirements, at least at a high level.
Then you’ll use this to build a preliminary picture of the services and service strategy that the company is following and gaining idea of the strategy it should be following to achieve good governance.
Finally, based on this understanding, you’ll be recommending a road map, with a picture of the first steps to be taken, and that will be the basis of your final report.
So, let’s pretend we’ve had that first meeting and go through the steps, based just on our knowledge and on the brief outline in the article. Remember that this is just an exercise and that you’d have to do a lot more proper work in any real situation!
Some firms trade on their own account and that is their main business. This requires a considerable amount of money and is, consequently, not so common. More usually, the trading entity is part of a larger financial organisation, a bank, a hedge fund, an insurance company or similar.
The particular trading that this article about, ‘high-frequency trading’ or ‘algorithmic trading’ is usually used for a small portion of the companies overall wealth with the aim of reducing risk and increasing return.
The reason that you’ve probably been asked to come in is that this form of trading, though fairly new and, in some ways, very dangerous, is surprisingly unprofitable. It costs a lot of money and doesn’t produce much return despite the high risk it involves. So little is lost getting your advice because they can try it without losing much and, if it works, they can gain enormously. This should help with some of your fears about the ethics of being involved – if you can reduce the risk, at least, you’ll be doing good for most of the stakeholders.
Let’s look at the stakeholders, and their main requirements, from the most remote to the closest to your direct client:
- The general public: less volatility, market growth based on genuine value of the underlying stock.
- The financial world: less volatility, market growth, reduced risk of short-term losses (avoiding another ‘flash crash’)
- The bank who’ll be paying you: reduced risk, increased profit, fewer missed opportunities, better governance of trading services
- The trading team: increased profit, reduced risk, beating the market, medium term stability
- The technical team: faster trades, better intelligence, better forecasting (short & long term), lower risk from software bugs, use of latest hardware technology, better algorithms, ability to change algorithms more quickly, more confidence that algorithms are accurately implemented.
As is often the case, the solutions (the software and hardware) change often, but the requirements are long term. You’ll need to understand what this company and trading team have as their particular requirements, and recommend setting up a requirements register, but the stakeholder analysis provides these:
- Improved governance
- Reduced market volatility
- Market growth
- Recognition of underlying stock value (should the trading team trade in their own Bank’s stock, possibly to its disadvantage, for example?)
- Reduced trading risk
- Increased trading speed
- Increased profit
- Improved forecasting
- Fewer bugs
Some of these appear contradictory, some are inter-related. From the point of view of ethical consultancy and the concern of most stakeholders, the key ones are – governance, reduced risk (which includes volatility, bugs & poor forecasting) and increased profit.
Services, risks issues, strategy and governance
It’s likely that, as the article describes it, the trading team is not thinking in terms of business services, but, rather, in terms of technology. It’s important to separate the various services to understand the investment cost and return of value per service, rather than just lumping them together. The actual services as perceived by the business would emerge in discussion, but a few obvious ones are:
- Fast buying service
- Fast selling service
- Forecasting service
- Risk analysis and management service (micro-risk)
- Algorithm design & deployment service
- Algorithm Service X – each algorithm has a cost/value profile based on risk, cost, stability, average & max/min of profit/loss in short, medium term
The issues are:
- Poor Governance: It’s pretty clear that the risks are very high because only the technical team understand the algorithms, the forecasts and the technology – so there is no way that effective governance can be in place at a board level because even the trading team can’t quantify the cost/value ratio for their services.
- Poor stock recognition: The emphasis on speed means that only a limited reference can be made to the medium and long-term underlying value of the stock. So software can easily be led into a bubble caused simply by instability, like the flash crash.
- No Control: The development and deployment, of algorithms, forced by the technology leads to poor control, high risk of error and poor forecasting
- Instability: The use of technology such as FPGAs means slow development times, high error rates and poor understanding of how the algorithms are actually working – reinforcing the lack of stability, high market volatility and poor governance. The banks don’t know what their machines are doing, and don’t know if they’re even delivering value for the bank, they may even be destroying value.
There are major short-term risks – of the worst sort, governance risks. No short-term solution is going to address them. It is, though, urgent to put in place some longer-term structures that can reduce the risk in the medium to long term.
- A service portfolio should be constructed for all business carried out by the trading teams, so that particular deployments of hardware and software can be understood in terms of the value/cost ratio, based on a genuine understanding of risk, volatility, forecasting accuracy and measured profit/loss ratios in particular markets
- Metrics need to be designed to measure the effectiveness and efficiency of these services
- A requirements register should be produced so that algorithm design and deployment can be tied to compliance, corporate strategy and policy as well as trading team objectives. This must be tied to the corporate risk register
- A plan must be produced to replace the high-risk development techniques being used. Firmware design and development of ASICs, GPUs and FPGAs is slow and error prone if carried out in low level logic design, assembler/microcode or unreliable languages such as C++. This should all be replaced by a corporate policy to use only Ada as this is reliable, documents properly, is designed for embedded systems of this nature, and is quicker to develop and faster (this is proven empirically) than assembler or microcode.
- Forecasting must be designed to incorporate board policy, longer-term views of the value of stock and to use the power of Ada to reduce volatility by enabling decisions to incorporate longer views.
- Staff need to be trained in service management to understand the service metaphor and to start to understand the contribution of services in terms of their contribution to business value, not simply to technical short-term gain (even if the algorithms themselves actually exploit this)
Do you agree with Peter’s recommendations?
For more advice on ITSM consulting check out Peter’s publications: