On time and on budget

January 23rd, 2012

I’ve just seen some of BBC1’s Panorama programme, “Taken for a ride?”, which is about the rises in train fares and the costs of railway infrastructure. Now, the segment I saw (the last fifteen minutes) consisted of the journalist talking to various figures in the rail industry about the cost of project work, and in particular the large overspends on some rail infrastructure projects. He kept asking “will the project be on time and on budget?” and using phrases like “is our money being wasted?”. Now, this is the standard stuff of all projects, so let’s examine this in more detail. What do we mean by “on time and on budget”?

Well, a typical project, be it the construction of a garden shed or a high-speed railway line, starts with a proposal. In that proposal, the promoter of the project sets out the benefits the project will bring, the risks associated with doing it, and their estimate of how long it will take and how much it will cost. Note the crucial word there – estimate. The proposal is scrutinised by some decision-maker, and (all being well) the project is authorised on the basis of the estimated budget and timescale contained within the proposal. Of course, as the project runs, some tasks will turn out to be easier than estimated, and others will be harder. So the project cost and timescale may change as a result. So, when we talk about the “ability of an organisation to deliver on time and on budget”, we really mean, to a large extent, their ability to estimate the project in advance.

Now, let’s assume for a moment that we’re going to put this project out to competitive tender, as all public sector projects are. A number of engineering firms will bid on the project, and that means that they have to spend their own time and money researching and estimating the cost and timescale of the project. There will, of course, be pressure on the engineers to produce a low bid, as the company wouldn’t want to lose the business to a cheaper competitor. There will also be pressure not to spend too much time doing detailed research into the cost of a project, as the cost of the engineers’ time is at the company’s own risk – if they lose the bid, all that effort will have been wasted. So the tendering process is likely to produce optimistic estimates of the project outcome. If a bidding company has done a project like this before, they will make a better estimate – but a rival with less experience is more likely to make a low bid, particularly if they need the business badly. So the tendering process needs to evaluate the companies’ bids clearly, and not just look at the price. There’s always temptation for companies to bid low to secure the project and then ask for more money once the project is beyond the point of no return.

Well, what about making the job a “fixed price contract”? This transfers all the risk from the buyer to the supplier. Great for the buyer? Well, everyone bidding on a fixed-price job will include an additional contingency for unforeseen circumstances. Depending on the nature of the project, that “risk premium” could be enormous. Most companies taking fixed-price work will insure themselves against it, so now you’re paying an insurance premium and asking an insurer to take the risk, for which they charge a considerable markup! The fixed-price job works well when the job is well understood, but for big, risky projects, the fixed prices become ridiculously high.

So, perhaps we should be more circumspect before we condemn projects that run late or “overbudget” – how well can you estimate?


Business hours and ‘housekeeper culture’

November 27th, 2011

An article in today’s Observer made mention of a service called Hubbub, started in Islington (of course! Honestly, the Guardian should change its name to the Islington Gazette…) which offers an online ordering and delivery service for buying food from local independent shops. In a similar tone is another article from the Guardian about the value of traditional markets in towns and cities. Now, I entirely buy into the argument that food purchased from independent retailers is often better quality, cheaper and more sustainable than that from supermarkets. I would dearly love to wander, wicker-basket in hand, around a lovely old covered market hall and buy my fruit, veg, bread, meat and other delights from jolly local traders who know and understand their produce. However, I don’t. Almost all the food that I buy comes from Sainsbury’s or Tesco. Why? Because they’re open at times when I can go shopping.

Now, I’m not working silly hours, either: I leave my house at 8am to be at my desk (12 miles away) at 8:45 – and I leave work at 5:30 and get home by 6:15. However, that effectively excludes me from dealing with all independent shops except at the weekend. Now, I could, of course, go and do my weekly shop in the city market, and perhaps visit Northrop’s the butchers’ in Mill Road on my way home. The market trades primarily in the morning – by mid afternoon the produce has been sold – and Mr Northrop doesn’t trade on Sundays (he is due a day off, after all). So that effectively means that Saturday morning has to be earmarked for food shopping, and of course there are lots of other people in the same boat as me, and so the city is heaving. This also limits what you can do with your weekend, and means you have to be out of bed on a Saturday morning before 10am…

So far, so much whingeing. However, I think there are a couple of wider points worth making.

Firstly, extending business hours into the evenings would massively increase the target audience of small shops, and would potentially allow them to compete more readily with big chains and online retailers. Of course, the shopkeepers themselves would need to take on extra staff to cover the extra hours, which would help with our present unemployment problem. The challenge would be to get a critical mass of shops to all open together. Cambridge supposedly has late night shopping on a Wednesday (meaning that the shops close at 7pm, woo, that’s late!) – but by no means all shops participate. For some businesses, a change in their work patterns would also be required. Bakeries, for instance, would need to keep baking through the day in order to have fresh bread in the evenings – in Portugal they do just that, and many people stop by the bakery to buy bread and cakes on their way home from work. Finally, by keeping the shops open, our town and city centres remain busy into the evening – which reduces crime and anti-social behaviour that can otherwise happen in deserted shopping streets after dark.
I’d be seriously in favour of a trial in which government under-wrote the costs of opening shops later in town centres, on the expectation that the retailers would boost their profits and the social benefits of a lively evening culture in towns would save money in other areas of spending. Managed well, I think it would be a great success.

Secondly, there’s a broader point about how we organise our society. We still seem to have a lingering social expectation that someone in the household will be not working during office hours. Courier companies and the Royal Mail attempt to deliver parcels during the day. Last week, National Grid (at four days notice) told me that they needed access to the house in order to replace the gas main. As soon as you bring children into the picture, the cost and hassle and issues associated with childcare, the desire to bond with your own offspring and the societal expectation of someone being at home, is it any wonder that so many parents -mostly women- choose to give up work?

In other countries shops trade late – why not here?


Missing the misguided bus

August 17th, 2011

I went on a training course today, and chose to take the train to the venue instead of driving. I arrived back at Cambridge station at 1740, and went in search of a bus. As I came out through the station entrance, I saw one of the new guided buses on the roundabout – it was snarling up the traffic as it was waiting for temporary traffic lights on the new access road that leads to the guideway. In my transport-fanboydom, I set off at speed down the footpath to the new bus stops for the guided busway, only to miss the said bus by about a minute. Bother. I looked at the timetable: this is peak hour operation, but the frequency on the southern part of the busway (which goes to Addenbrooke’s Hospital and Trumpington) is only one bus every 20 minutes. I walked back to the main bus stops and checked the timetable there. Normal, non-guided buses run to Addenbrookes every 5 minutes, and to Trumpington every 10 minutes. This begs the question: what is the southern section of the guided busway actually for?

Let’s have a look at the timetables. Only one of the four busway routes uses the southern section at all: route A, operated by Stagecoach. This operates one bus every 20 minutes right through the day, stopping abruptly in the early evening: the last buses are at 1830 from Cambridge and 1839 from Trumpington. A note hidden at the bottom of the timetable shows that there are an extra hour’s worth of bus services between Trumpington and Addenbrooke’s after the main service finishes.

The journey times on route A are as follows:

Trumpington – Addenbrookes: 6 mins
Trumpington – Rail station: 13 mins
Trumpington – City centre: 22 mins
Addenbrooke’s – Rail station: 7 mins
Addenbrooke’s – City centre: 16 mins.

Okay, so how does this compare with the conventional bus services? The guided buses have a dedicated route from the railway station to Addenbrookes and Trumpington, but going into the city they are on the road.

Let’s take the Trumpington – City case first. On this route, the guided buses compete directly with the Park & Ride shuttle bus, which makes no intermediate stops between the P&R site and the city. The P&R buses run every ten minutes (twice as often as the Route A buses) and take 16 minutes for the journey. So they are 5 minutes faster than the guided buses, despite being on the road the whole way! As they run twice as frequently, the average journey time for a passenger arriving at the P&R site is 21 minutes (5 mins avg wait + 16 mins journey) for the P&R buses and 32 minutes (10 mins avg wait + 22 mins journey) for the guided bus.

How about Addenbrookes – Railway station? This bit is on the guideway, and so should be faster. And so it is – 7 minutes for the guided bus, as against 10 minutes for the conventional service. However, as the guided bus is only every 20 minutes, whereas the Citi 1, Citi 7 and Citi 8 services combine to provide a bus every five minutes. If you walk out of Addenbrookes at some randomly-determined time and go to look for a bus, you would on average wait 2.5 minutes for a conventional bus, followed by a 10 minute journey time – i.e. 12.5 minutes total journey. However, your wait time for a guided bus will be an average of 10 minutes, followed by a journey time of 7 minutes, making 17 minutes average total journey time. So the guided bus is not much use for going to the station from Addenbrooke’s. The same argument applies to taking the bus to the city from Addenbrooke’s.

Okay, so what about Trumpington – Addenbrooke’s? Maybe the idea of the southern section is to act as a P&R service for the hospital? Well, firstly the every-20-minutes frequency is not exactly encouraging. Secondly, parking at Addenbrooke’s for hospital visitors is not astronomical: £3.50 for 2 hours costs the same as a Dayrider ticket on the guided bus. Thirdly, Addenbrooke’s has a nice new access road that connects it to Trumpington P&R site and the M11, so if a hospital P&R service was required, there was no real need to build a fancy new busway – the buses could just as easily use the new access road.

So, to summarise: the guided buses only offer faster journey times on certain parts of their route (and even then we’re only saving three minutes) but their low frequency means that most people will be better served by the conventional bus service. Providing a P&R service between Trumpington and Addenbrooke’s could have been easily done without the guideway. So what exactly is the southern section for? I still don’t know…


Plusbus

June 26th, 2009

If you’re a regular passenger on the railways, you’ll probably have seen adverts for PlusBus, which is a scheme that allows you to purchase discounted bus tickets if you need to continue your rail journey by bus. It’s a great idea but has suffered somewhat from implementation issues. PlusBus is run by a consortium of bus companies (most of whom also own train companies) but as a consequence the railway staff aren’t always very clued up about it.

The biggest problem with it is actually buying the ticket in the first place. Most self-service ticket machines can’t currently sell a PlusBus ticket, which isn’t much help if you travel from an unmanned station. When I lived out at Waterbeach I would buy an ordinary return ticket to Cambridge and then upgrade to PlusBus at Cambridge station, but if there was a huge queue at the booking office it wasn’t worth it. My worst PlusBus experience was one time I tried to go to Leicester. I wanted to buy a ticket in advance, with both Cambridge and Leicester Plusbus upgrades, so that when the tickets arrived in the post I could get the bus to Cambridge station, train to Leicester and the bus to the university. At that stage no online ticketing company would sell you a PlusBus. I had to ring about five different people at different train companies before I finally found someone who could sell me a ticket.

Today, PlusBus has come back into my attention. I had a hire car to drop off this morning at Hertz, which is by the railway station, and then needed to get to work on the bus. Because I have to take two buses to do this, I’d need a Stagecoach Dayrider ticket, which now costs £3.30. However, since I’m going to London this evening, I went into the station and asked the booking clerk for a return to London with a Cambridge PlusBus. He said I could only have a PlusBus for arriving in Cambridge, and not for departing from it. I told him this wasn’t true, and I’d done it before. He went off and referred to a colleague, before returning to sell me the ticket. After I’d paid, I realised that he’d forgotten to give me a railcard discount on my PlusBus, so I’d paid £2.70 for it instead of £1.80. Oh well, but still better than £3.30! Interestingly, the PlusBus ticket covers a much bigger area than the Dayrider (more or less equivalent to the £5.00 Dayrider Plus ticket) so if you have a railcard, it’s actually cheaper to buy a single ticket to Shelford (the next station south of Cambridge) and a PlusBus (total cost £3.30) than it is to get a Dayrider Plus.

One other thing – each PlusBus ticket is only valid for 24 hours, so if you buy a period return and want to use the buses on both days you travel you need to buy two PlusBus tickets, one for each day.

However, all this prompted me to look at PlusBus’s website again, and to discover that the tickets can finally be bought online. Unfortunately, only two providers can do it. One is Qjump who charge a booking fee depending on how you pay and how you collect your tickets, details here. The other is First Transpennine Express, who operate trains in northern England, but whose website is powered by Qjump. Happily, Transpennine’s website will sell you a ticket for any part of the country and doesn’t charge a booking fee. I shall be using them a lot more in the future! Nothing like making it easy, is there?


Experiences with Simulink’s Communications Toolbox

April 4th, 2008

Simulink is a companion to Matlab which purports to be a time-flow simulation package. This post is going to get quite technical, so if you’re not into either communications system design or Simulink in general, look away now.

I’ve used Simulink before, in my third year at York, but that was following a well-structured lab script where the lecturer had partially set up the models in advance. This time I’ve been doing a course (I’m obliged to do one of Leicester’s engineering MSc modules in my first year) simulating communications systems in Simulink, and the lab script is much vaguer. I’ve spent a lot of time banging my head against a brick wall in Simulink, and here are some important gotchas that might save other people going through the same level of auto-cranial damage.

Read the rest of this entry »


The MathWorks, Inc are a Bad Railway

February 11th, 2008

You may have come across MATLAB, the matrix-manipulating maths package that’s widely used by engineering academics. It’s published by an American firm called The MathWorks, whose slogan is “Accelerating the pace of engineering and science”. This appears to be, in the immortal words of Douglas Adams, a load of dingo’s kidneys. I’ve been trying to buy a copy of their software (MATLAB, Simulink and the various Communications add-ons for them) since before Christmas, and they’ve now wasted over a month of my time in failing to supply it.

The complicating factor is that MATLAB is extraordinarily expensive. The list price for my combination of options, as a single-user licence (that is, a licence for just me to use it on my computer) is over £6000! However, MathWorks have always given huge discounts to universities, which is why it’s so widely used in academia. BAS have been negotiating with MathWorks for about a year to get the academic price on the software (which would seem reasonable given that we’re a government-funded academic research institution) and just before Christmas I was told that the team had negotiated for a significant discount. I promptly raised a purchase order before the Christmas holidays set in, then came back to find nothing had happened. I’ve spent all of January chasing Steve (the purchasing officer) who has in turn been chasing MathWorks by phone and email and even refaxed the purchase order to them, and still no answer. Today I return to work after being on holiday for a week (Skiing in the French Alps, marvellous!) and find that a) nothing has happened and b) Steve is on holiday. I get a copy of the correspondence from one of Steve’s colleagues and chase the firm myself. Their salesman and I play voicemail tennis this morning and I eventually get a message from him saying that we need an invoice number to qualify for the discount. I go and see David, the Head of Engineering, who’s been doing the negotiating. He calls the salesman, who promises to email us more details of what hoops we need to jump through.
The saga continues.

In the mean time, I would like to say that The MathWorks, Inc, are a Bad Railway. If anyone else is planning on buying their products, I would seriously reconsider it!


Urban bus rant

April 24th, 2007

I´ve taken a lot of urban buses in the last couple of weeks, and this has resurrected an old rant aimed at bus companies and local authorities.

So, you arrive in a strange town by any mode of public transport. Nine times out of ten you will arrive at an airport/railway station/bus station that is some distance from your onward destination. The city has buses, clearly, but where do they leave from? How often? How much is the fare? Will the driver need exact change? This information is never available in an obvious form in the terminal. This is especially galling when you´re at a terminal for long-distance buses which is (as in Bariloche) 3km outside the town centre!
Similarly, in lots of cities, bus stops are unmarked. Yes, Bariloche and Rio Gallegos I´m looking at you especially. Even in British cities, you usually just get the sign saying “Request Stop” and nothing else. You´re expected to know where the bus goes and how often. In my village there were bus stops for a service that only ran during school terms – how were you expected to know that? If you´re lucky enough to see buses you have to resort to stopping every one that looks promising and asking the driver if he goes where you want. If you´re lucky, he´ll tell you which bus you want in a tone that doesn´t imply that you´re a bloody ignorant gringo/southerner/northerner/tourist.
It gets worse. In most towns here in South America the streets are in a grid pattern with each street carrying traffic in one direction only. So if you get off a bus somewhere, the return leg of the same bus route will be on one of the other two parallel streets, but which one? You don´t know because there aren´t any marked bus stops!
British towns are no better. London actually gets it right, with its very handy Underground-style route maps at every bus stop showing where the buses go and at roughly what frequency. But other cities are lamentable. Cambridge in particular has one of the most actively user-hostile bus systems in the world, particularly if you arrive at the railway station which is, you´ve guessed, over a mile from the city centre.
Bus companies are also fond of highly circuitous bus routes that take you several blocks in the wrong direction before doing a 180 degree turn (Bariloche again) or totally bizaare topologies (like the figure-of-eight bus route that used to run between York University and the Monks Cross shopping centre). The fare structures are often non-obvious and priced to a bizaare level of precision (1.15gbp, say, or AR$1.30) rather than just a flat fare to anywhere on the route (London). This is particularly frustrating where the bus drivers are either snotty about giving change (Argentina generally – there seems to be a national shortage of small change and two peso banknotes) or are panelled in behind security glass and only take exact money (Birmingham and Coventry).
None of this is rocket science, but people won´t take buses unless they either do it every day or they know how it works. Maybe the taxi drivers are conspiring behind the scenes to ensure that tourists find buses impossible to use!