Tuesday, March 27, 2007

What if hedge funds ran outsourcing companies?

Barely a day goes by without a new story linking a trio of private equity firms to increasingly implausible bids or, at least, bids that would have looked implausible as little as a year ago but now look increasingly viable. Retail, a sector with notoriously thin margins and comprehensive execution risk (cf Sainsbury and Marks and Spencer in recent years), has been a particular favourite recently. Now, rumours and facts abound about various IT outsource companies coming under scrutiny. Another sector where margins are thinner than thin (size zero?) and the execution risk often strikingly large. What would happen if the hedge funds applied their "2 and 20" rule to the outsourcing deals that they win? Two and twenty is where the fund manager takes a flat 2% of assets under management and then 20% of any profits, at a gross level. The actual amounts vary - some funds charge as high as 50%, but they deliver, in return, pretty stellar gains year in year out. So imagine a company, Joe F Smith & co, signs over its back office services to TwoAndTwenty Capital Partners. JFS pays £100 million per year to operate the services in-house. TATCP offers to do it for less, but wants £2 million flat fee every year as well as the £100 million. If TATCP can reduce the cost of operation, they get 20% of that reduction and the customer, JFS, gets 80%. Is that a good deal for JFS? They pay £102 million for what used to cost them £100 million. If the cost falls to, say £80 million, they pay £86 million (£80 million plus £2 million flat fee plus £4 million as the 20% of saves). TATCP gets £6 million versus its original £2 million. What if costs go up? Strictly, under hedge fund terms, the clients carry the loss, the hedge fund gets its 2% performance fees but doesn't get any upside (the "20" disappears). Further, in following years, the hedge fund has to earn back the lost money before its allowed to take any of the "20". This keeps the fund manager keenly focused on performance and, also, on stability (i.e. low volatility) - making £100 million one year and losing £100 million the next year makes for zero performance earnings over two years and the problem of having to gain the money back the following year before any further upside is received. In a hedge fund world, there is unlimited upside of course - returns could be 100%, 200% or even 1000% and so performance fees can rapidly become enormous. On the cost side of the equation, it's unlikely that services will be delivered for less than zero (I'm not yet sure there's room for advertising supported outsourced IT), so the returns will be capped. I wonder what the balance of 2 and 20 equivalent would have to be for this model to work? If it were 3 and 30 or 5 and 50 would a company take it up? There's always the risk that there'll be a long argument about accountability for cost increases that would defeat the point - the "hedge fund manager" is, after all, not in control of many of the variables unlike when they're trading and the choice of position is entirely theirs, even if they're unable to directly influence the market in most cases. Is this a better model than the glide path that is standard in many contracts where the cost goes lower and lower every year but where change control costs are always added to the top at the original margin? Of course, there are those who say you should never outsource ... but given that companies and governments are going to keep doing it and given, especially, that the big research companies say that somewhere near 2/3rds of outsource deals are significantly renegotiated or exited before their full term has run, there ought to be a better model from the outset that makes things more achievable. Applying this approach presupposes that the reason for the renegotiations or early exits are related to whether the supplier can make money or the customer sees itself getting value for the price paid - there are probably a dozen other reasons, but I'd expect this pricing/value point to be the single largest reason.

5 comments:

  1. Andy F-W12:43 am

    Someone from the cabinet office today said you earnt £2M from e-envoy.

    Is this true and what did you give the tax payer for that slab?

    ReplyDelete
  2. Anonymous3:57 pm

    CBI.com magazine interviews several prominent business leaders and government ministers, with a lie detector, a waterboard, a pair of pliers and a blowtorch, and asks the question, "Why do you outsource?"

    -------------- Ivan Ornérée Surcha MP repsonds : --------------------
    "Well it causes redistribution of wealth.
    If we outsource to a big company, we can have lots of secret deals where they open an office on the waste ground of an old ICI plant in the Middlesbrough slums, employing all the plebs who would vote us out if they were unemployed. It doesn't matter, per sé, that the project ever achieves anything or even works, the only important thing is that the government isn't embarassed by the employment figures, and if the project happens to be a success it's announced before a general election, and if a failure it's announced after we've lost the general election. Speaking of elections, I must see my researcher this afternoon.

    As a second point, if we didn't outsource, who would employ us when we left office?

    Obviously, we don't explain it that way, we explain it in terms of garnering the support of the whole community, and so on, but the reality is that we can push all kinds of social responsibility onto a big company, and they're given so much money for the job, I mean just look at the Child Support Agency, Microsoft Dynamics CRM, or even outlook, did all the things they needed but for just £500 Million we can bring the Northeast forward into the 20th century, that they can afford to employ all the care in the community types we would otherwise have to find secure accomodation for.

    As well as this, once there's one or two big companies handing the work, if we find say, a large amount of unemployment in a particular social group, say single mothers, we can create a new law or two to ensure they get a job. This technique doesn't work for small companies. For instance, the right to jobsharing, combined with any type of box ticking legislation, say equal rights, means a big company such as EDS or someone, has to prove they don't discriminate against the uncompetitive, and thus they have to employ diversities directors, and compliance staff. What else would a woman with a social studies degree from exeter do if we didn't create a market for her to work in? She'd be working in McBurgerland, that's what, and she'd vote us out I'll tell you what, eh? This is government after all, we have to look after everybody, not just the capable.

    Besides, empowering the weak has an additional advantage, in that we get to give women their own money, it's only a small step from giving Pakistani women their rights to their own money to the collapse of many a patriarchal religion eh? Can I rely on your vote?"


    --------Sir Withington Smythe, Head of Automata Systems, replies ---------

    "Why do we outsource? Well it's quite simple really, we've got a whole stockmarket out there baying for instant profits, and my job depends on it. If I say to the analysts twice a year that we've setup a large IT project, everyone thinks "They've never done a big job before!!", and they panic. Before you know it, the share price is down. If however, I say, we're getting -Insert Very Famous Consultancy Here- to do the work, they think, "Well we don't know anything about IT, but at least the CEO is prepared to admit he doesn't either. I mean just look at Aunty Pat, complaining about it being difficult to deal with all those database thingies, it just goes to show you see, inability to understand goes right to the very top. Besides, noone will ever hear if PWCMPG screw it up, because they'll settle out of court with a confidentiality agreement, and give me a job when I've left this place. There's no end of people happy they've had a board room table to fall back on."

    "There's a second problem, these IT people cost a hell of lot of money, and whilst I've no problem giving a six figure salary to old Billy Rothschild-Devere I fagged with at Harrow, but for some kid who is just out of school and grew up in an engineering yard, well they just won't be able to return the favour will they? I mean they don't even have a Hice in Gloucester do they? After all that's how I got started on the floor, noone's going to buy stock from a coalminer are they, but if they know my dad. The additional problem with these chaps is that they don't want to move into management, they want to stay technical. As for all the people who have social skills that you'd like to promote, the problem with these chaps and chapesses is that they don't have a ruddy clue about what happens inside a computer."

    "The final problem is that time has passed, and you've just got so many MBAs coming out of Harvard, Insead, LBS etc, that it's not like the old days anymore. There used to be a couple of hundred MBAs, but now they're like flies at polo match - and all of them want to cut costs, so they treat the staff like commodities. This was fine until the staff worked out how they were considered, so now everyone's just trying to guard the old canal. This is compounded by the fact that every young management pup has to justify his existence by suggesting that we do something else. I mean, noone wants to say, "It works. It's pretty good." because if they do, why do we need them, eh?"

    "Finally, there's one good thing about outsourcing. It allows the business to change. All business needs change, it helps us cycle the staff, otherwise you get these care in the community types building up. I'm not saying they shouldn't have a job, but I'd just rather they were employed somewhere else."

    "Of course the actual end decision to outsource has to be supported by some genuine business reasons, one of them is cost. There's a world of difference between cost and short term cost, just ask the Tessa or Seb, or Enron. Most businesses and most market analysts don't want the real cost of IT on their bottom line. The other is lack of skilled staff.

    Often it happens by accident. What happens is the finance director's walking around the office, and he notices that the 25 year old boffin's wearing dark glasses, because the previous night he was out in the bar round the corner in Wardour street, snorting cocaine, and flirting in flagrante delicitio, with some young girl. He thinks this "Just isn't right, I couldn't afford to do that in my day, and so engineers the removal of the CIO, to a new chap with an accountant background who lives in Sussex. He then immediately relocates all the staff to Croydon or somewhere nearer home, and then the exodus starts. Young chaps simply want wein, frauen und gesangen I guess.

    So then the CIO's starts a recruitment drive, but noone wants to work in Croydon, so they can't retain staff, and pretty soon all the people who can leave because of their ability, have left. At this point outsourcing the whole job to hedge the bets, seems like a pretty viable option."

    -------------------------------------------------------------------------------

    "Applying this approach presupposes that the reason for the renegotiations or early exits are related to whether the supplier can make money or the customer sees itself getting value for the price paid - there are probably a dozen other reasons, but I'd expect this pricing/value point to be the single largest reason"

    In the words of the great Paul Newman, "You betcha!"

    I agree with this.

    As with most of my ideas, I have yet another set of boring empirical observations as a starting point, followed by a set of conclusions based on inductive reasoning, and then a simple reversal of the point of view, to make it sound different.

    The observations followed by sub conclusions.

    1. Most large IT projects are also business process re-engineering projects.
    2. Most people know what they like, and you can lead a whore to Walter.
    3. Mensa says you are born clever. Most non-pop Psychologists say your personality is fixed at birth, barring extreme trauma. My cousin says almost everything is fixed at birth and he's a very clever chap. (He's a Phd geneticist, and he makes me feel positively insecure at times, despite having dated Phd Physicists who haven't.)
    4. In a lifetime of working in Software, there's a direct (and decidely non linear) correlation between cleverness IQ and motivation (collectively called g, cognative ability,) and the respective achievement. I'd go further than this and say in my career, on maybe 15 projects, the average team of 25 people will have a couple or three lads in it, that provide 90% of the output between them,
    In fact this isn't even just in Software, just look at Henry Ford, Einstein, Bill Gates, Ellison, etc. All dropouts AFAIK but with extreme talent and motivation. Qualifications have nothing to do with it. It's talent, and motivation.
    5. From the experience of watching the builders working on my house, if I let them get on with it and come back four months later when it is finished, I get what they want. If I go home every day at lunchtime, with pound notes, I get what I want.
    6. In order to get a fireman, or a hairdresser or a mechanic, you start with a man and train him. In order to get the Royal Marines, or the SAS, or World record holders, you start with the world and select out all those who can't do it.
    7. Big jobs are more likely to fail than small ones.
    8. A guy I bought a house with, who was a gambling addict, once correctly pointed out when I pointed out he'd lost 300 quid in a fruit machine, "If I put in another 100 I'm likely to get 250 back."

    So some conclusions may be, that

    a. You can't trust someone to build something you like, or does what you want, you have to keep your eye on it, with frequent regularity.
    b. You won't be able to trust anyone oneday if they know you're going to become dependent on them, because there's no easy get out or transfer clause, see elastic and inelastic demand.
    c. You can't train someone who doesn't have natural technical ability but otherwise has management aspirations to have techical ability. (Similarly you can't teach management genius either.) All you can do is select out those who don't have both.
    [You'd obviously need to ask a psychologist about this, but I suspect that since I've observed loads of clever techies refusing management, that many people go into management because they know they can't do anything interesting. Management is much more subjective than delivery after all.]

    Thus we now look at some of the software development masterstrokes in history.

    There are two that stand out. One is the development of C/Unix. This was a massively parallel collaboration, where everyone spoke out, and it happened in small steps.

    The second was far more brilliant, the development of the Ada programming language, (I'm not talking about Ada 95, or even Ada 2005, which while good - all credit Robert, but were not the equal of Ada 83. Mil Std 1815, developed by the US department of defence.)

    So there was this Frenchman, I assume, who just happened to be technically brilliant, but he also was a business genius as well. His name was Jean Ichbiah.

    What he did was get _SEVERAL_ suppliers to come up with their ideas, and evaluated them, combined them and resubmitted for advancement. They then resubmitted the next step, he combined the best ideas, and asked the companies to come up with improvements. You'll note here that more than one company was operating to achieve the same result. It was redundancy of development in small steps, none of the companies could screw over the US dod, and they all got to start the next step from the same position, a bit like the golfing game Texas scramble.

    After about four or five iterations, the Ada programming language was born, and it worked so well it's still used, despite being despised by the C crowd who were young peace love dope kids from Berkeley et al.

    He did it all right. He didn't make a single mistake. Unfortunately, there aren't many of him out there.

    So the key ingredients are.
    1. Small step development, with multiple suppliers delivering a step in parallel. This avoids tie-ins.
    2. A small integration team inhouse, and staffed by extreme techies with an MBA, or MBAs with a degree in Computational Physics, who review and integrate the advancing prototype, and get each of the parallel streams back into convergence for review my the people on the shop floor. Loads of techies know years in advance that something isn't going to work but they aren't heard, because they haven't got the business skills.
    3. The resubmittal process for the next step of advancement.

    Things never get out of hand.

    The problem is of course that sometimes ministers want things to fail, just look at I*L P*th**y for instance. If various parts of it had worked, then it would have shown up the B*n*i*s A*e*c*, and that would have been embarassing wouldn't it? So the overwhelming priority is do you actually want the project to succeed.

    No that I'm saying politicians are lying slags at all, because I resolved to be more like a friend of mine this year, and to stop being negative.

    I.

    ReplyDelete
  3. Anonymous5:30 pm

    "Applying this approach presupposes ... but I'd expect this pricing/value point to be the single largest reason."

    I agree with this assertion, but I feel I've been unlucky. I've heard of loads of cases just like this, however in my own career, most cases come from...

    1. Consultancy gets in.
    2. Talks the job up.
    3. Convinces the client they need all kinds of other functionality.
    4. This behaviour stops the original idea working.

    This has analogues in real life. In my physics class at sixth form, my teacher had this huge roller on the wall. He suggested that we measured it's density or something, which I concluded the best thing to do was to wrap a rope around it, with the heavy weight pulling the rope round until it hit the ground, the time taken and the height gave us the acceleration and thus given the angular acceleration we could simply do a reverse calculation from the shape, the integration of which would give us its mass. From there it's another simple calculation to work out the volume, thus giving the density.

    Alas someone else then suggested an enhancement, that they measure the time from one place to another, in order to rule out the standing friction. This of course meant there were too many unknowns and it all went pearshaped. (Actually this is slightly innacurate reporting on my behalf, because even in my state of age related retardation I can still see it's possible. The only explanation is that I've forgotten what the original question was.)

    Anyway, you get the point.

    Consultants seeking scope creep, so more bums on seats, so more money, eventually causing impossibility of delivery.


    Ian

    ReplyDelete
  4. re: "Someone from the cabinet office today said you earnt £2M from e-envoy. Is this true and what did you give the tax payer for that slab?"

    Only 2 million quid? I'm disappointed the rumours weren't for far more. I figured I was good for 5 million at least. Didn't they mention additional perks I was given, like MPs? The house in the country (slightly larger than chequers), the chauffeur driven car at my beck and call and the helicopter available at battersea heliport whenever I wanted it? Or the private jet and the villa just outside Nice?

    If you're expecting a bob kiley-style "not much" in response to what did I do ... no can do, but here's a runthrough of what did go on:

    The team that I ran, the people that I worked with and for were responsible for a broad range of things: the government gateway (10 million registered users), the original and revised ukonline sites, directgov (2 million users a month), dh.gov.uk (1.2 million users a month), the rescue of the 1901 census site after its disastrous launch (bringing regular and pretty good revenue to government as far as I know), the first advertising engine that let government sites advertise each other's services, the first pan-government anti-virus service (with ogc), pan-government search available to all government sites, the first sale of government developed software to an overseas nation, the criminal justice exchange and secure mail services, development of the gsi (with ogc), advice and guidance to many departments including gate reviews for some of the biggest IT projects in the uk. I could go on. what have you done lately?

    ReplyDelete
  5. Michael S2:55 pm

    there you go andy f-w. Ball back in your court I think.

    ReplyDelete