Banks get tough on performance 12 Dec 2006 Hundreds of staff face job cuts and bonuses are threatened in ‘preparation for downturn’ Investment banks are taking their most aggressive approach to performance and bonuses for four years and are planning hundreds of job cuts over the coming months. Banks will cut more staff than at any time since 2002 as part of their annual performance reviews, according to senior bankers. At the same time, they will slash bonuses for underperforming staff to encourage them to leave, while lavishing bigger bonuses on their best performers. Headhunters and senior bankers also said a number of big banks have drawn up contingency plans for further cuts if there is a market downturn in the second half of next year. The existing cuts will give them greater flexibility to invest in high growth areas. Last week, Dresdner Kleinwort, HSBC and Royal Bank of Canada set the tone for the next three months by announcing more than 300 job cuts between them just before bonuses were announced. Dresdner Kleinwort last week instigated an aggressive programme of cuts that could see as many as 7% of its staff lose their jobs. The bank employs 2,500 people in London. The cuts, reported first by Financial NewsOnline, have seen dozens of managing directors lose their jobs. A source close to the bank said the changes were a reflection of a more aggressive performance review process and the “flip-side” to a new bonus scheme announced in July to link bonuses more closely to profits made by individuals. HSBC and Royal Bank of Canada have made deep cuts in fixed income, where trading conditions are expected to be tougher next year. Most banks cut between 1% and 5% of their staff each year as part of performance reviews. In the past few years the cuts have been at the lower end of this range, as banks struggle to keep up with high volumes of business. The head of one investment bank in London said: “We are applying greater economic discipline to our costs this year than for several years. We can only afford to pay our best staff competitively if we don’t waste money on those who do not deliver. For the past few years, the industry has been growing rapidly. While next year looks like it will remain strong for investment banking, there are concerns over the interest rate environment, trading profits and levels of leverage.” One big US bank recently held an offsite meeting for its senior directors, where one of the sessions was called “Preparing for the Downturn”. Banks are grappling with the problem of retaining their best staff in what is likely to be a record year for European investment banking, while keeping control of compensation costs with an uncertain market outlook. One banker said: “We are trying to differentiate between those that have really performed and those that haven’t. A greater number of people will get no bonus this year in a move that will encourage them to leave. A small number of top performers will get a big bonus.” One fixed-income headhunter said: “The job cuts could prove to be the first sign that banks are tightening their belts ahead of a downturn. We haven’t seen a wave of industry-wide job cuts since 2003, but there could be one just around the corner.” The tough approach follows a record year for investment banks. This week, Goldman Sachs will kick off the US banking reporting season with record results, but analysts are predicting a 6% fall in earnings per share at the bank next year. Some think revenues and profits could fall across the industry in 2007 for the first time in five years. The Securities Industry Association said in a recent report that profits at Wall Street firms could fall by 22% to $20bn (€15bn) next year due to margin compression, slower economic growth and a fall in activity in the primary and secondary markets. Bankers are reporting healthy deal pipelines for the next six months but are unable to predict beyond that, as they fear excessive leverage on private equity deals could lead to greater defaults and a market correction. The chief executive of one European investment bank said: “We are increasingly concerned about the high levels of leverage and valuations of buyouts. We do not think these levels are sustainable.” Nearly three quarters of people at a Financial News conference on leveraged finance last week said they expected a downturn and an increase in default rates in next year.

Pretty much my view, although not as alarmist below, will take a little time for the effects to mean firing people - but that will be around 2 years away. Next year will be announced as "A year to Consolidate our position", and there will be a lot of guff about competitive markets, and having work smarter etc.

I was thinking of moving to UK... now i have think again... I think jobs opening will be more in emerging countries... then Europe n US...

QuoteOriginally posted by: rmax... Next year will be announced as and having work smarter etc.If I may reduce the scope and give my 2 cents on the software aspects (and what areas of life are not software these days? )...In my experience during the last years: in the event of a slowdown people work even smarter and this means that time-to-market is reduced, quality improved etc. etc. etc. In many cases this translates to the ability - either individually or as a group - to manage the lifecycle of software projects, including estimation, planning and scheduling and monotoring. And software economics can also play a major role. Raw programming talent augmented by s/w management skills.What I have seen is under the current scenario is the trend to maintain and improve systems and processes. This is like a software form of Kaizenhere is a book on continuous improvement. Many tips and practical guidelines (I suppose it is same as 'get it working, then get it right, then get it optimised') hereOne project that I did once was to refactor and redesign a large and 'deep' C++ class hierarchy to make it more reusable and robust. Oh yes, it was documented as well in UML It is a non-trivioal task especially if the inheritance tree is anyway 'complicated'

QuoteOriginally posted by: CuchulainnQuoteOriginally posted by: rmax... Next year will be announced as and having work smarter etc.If I may reduce the scope and give my 2 cents ...Is that 0.02 EUR or 0.02 USD. Cable rate isn't so good nowadays I have seen numerous projects trying to change processes and make then work smarter, but it ends up with millions of wasted dollars. System architecture on a Darwinian process is required. Improve current processes and systems until they get consumed by new ones. Don't try and build big systems unless it is your core-compentency.

double

QuoteSystem architecture on a Darwinian process is required. Improve current processes and systems until they get consumed by new ones. Don't try and build big systems unless it is your core-compentency. Common sense and the correct approach.I asked one of my developers once "how is it going"; answer was "we are 90% finished, the UI looks great, and we are now starting on the routing algorithms" Unfortunately, the product was due to be shipped 2 months later!!Where did it go wrong? It looked great all along and then shock.

Agreed common issue. The classic Word for Windows 1.0 scheduling issue (as taken from McConnel RAD book). I have extracted a couple of dates to compare:Report......................Est Ship......................Est Days......................Actual DaysDate.........................Date...........................To Ship........................To Ship=================================================================Sep-84....................Sep-85...........................365...........................1887Jun-85.....................Jul-86.............................395...........................1614Jan-88.....................Jun-88............................152............................670Jun-89.....................Sep-89...........................92.............................153Jul-89......................Oct-89............................92..............................123Aug-89.....................Nov-89..........................92..............................92When I first read it, it was one of the most useful tables ever as it changed my behavior as a Project manager

QuoteOriginally posted by: rmaxAgreed common issue. The classic Word for Windows 1.0 scheduling issue (as taken from McConnel RAD book). I have extracted a couple of dates to compare:SEE rmax datesWhen I first read it, it was one of the most useful tables ever as it changed my behavior as a Project managerThis looks like this team got better and better in estimation?For one-man projects I ask people to do as follows (because they are over-optimistic):a = optimistic durationb = realisticc = pessimisticthen expected = (a + 4b + c)/6This is for each activity in project; we do a risk analysis at the same time; developer must explain why he says it take 23 days work.For the total project, these estimations work out nice.

Not sure the they got better - I read it that they kept asking how long to go, and they came up with the old is only about another 3 months work to go each time. I think the work slowly converged to their estimation and not to their estimate to the work.I like the weighting function - do you do any backtesting to see how accurate it is?

QuoteI like the weighting function - do you do any backtesting to see how accurate it is?Yes, because the project here is fixed price, the estimates are sent to the customer who signs it off. In our experience the algorithms are tricky to estimate (especially if the precedence ("have we done it before?") factor is low) but then the worst case estimate tends to be higher (flattened curve).What is good is that risk tends to shared between sponsor and dev team. Statistically this method is accurate 68% of the time, especially when averaged enough (20!) activities in a project. Here is a story on this estimation technique Pert & Gert All practical stuff; it attempt to get the project up and running on time. P.S.The very fact that you sit down and think about how long the project takes is good; forces you to realise why you underestimate in the past.

I favour MS Project as a way of calculating deadlines.The problem is gathering the numbers objectively, and this helps a lot.It's inmteresting that when you get people to break things down into tasks and guess how long they will take, they keep wanting to say "I know it takes 4 days to do this, and 10 to do that, 6 to test and 2 to sort out the bugs, but I reckon it will take 15 days to do the lot". In my invariant experience, the operator "+" when applied to two timescales rarely delivers a number equal to their sums, but is generally longer.I favour the sprint/grind model of programming.Lots of work is bulky, but low variance in it's timescales, and you can sprint through this.Some problems are really hard which means both that they take up chunks of time, but also that the variance is obviously greater.When you ask a programmer for an estimate, it's very easy to make the mistake of saying in effect."give me a list of the problems you can't predict, and precise times to solve problems whose nature and number is as yet undefined"I try to extract from people some notion of their view of the obstacles, but am mindful that knowing what the problem is, comprises 2/3 of solving it.It's also a mistake to set the timescales before you've questioned the ream who are going to build it.It is also a mistake to let "business analysts" into your building.

QuoteIt's inmteresting that when you get people to break things down into tasks and guess how long they will take, they keep wanting to say "I know it takes 4 days to do this, and 10 to do that, 6 to test and 2 to sort out the bugs, but I reckon it will take 15 days to do the lot". In my invariant experience, the operator "+" when applied to two timescales rarely delivers a number equal to their sums, but is generally longer.I think that there are number of reasons for this discrepancy, which are well-documented in the literature (See Boehm's tome)- developers wish to please (4 days to complete soundd nicer than 10.5 days)- incomplete recall (people forget admin, preparation part, meetings, research)- software is compact, right? so it should not take so longAnd a big one is- What is the critical path <==> get it working.

My algorithm is much simplier.... 2 monthsIt is only a question of how many rent-a-bodies you throw at the problem!!

QuoteOriginally posted by: LTrainMy algorithm is much simplier.... 2 monthsIt is only a question of how many rent-a-bodies you throw at the problem!!Interesting but this is not my experience. This approach you advocate does not take risk into account. So I increase the numbers even more and we do it in a month Some successful projects start off with a (very) small team that does the architecture and basic working model and then when things are stable the 'posse' comes on board (see Boehm "Software Economics") Boehm and Spiral model