Is the whole world going downhill, or is it just me?

One of the key skills of an analyst, when noticing some important ups or downs in the results of a competitor, is trying to decide whether these anomalies pertain just to that vendor, to a particular pocket or two in the market, or to the entire market.

Take the current economic uncertainty. The key question everyone wants answered is: how is it affecting IT spending?

The analyst firms such as Gartner and IDC conduct market research to find out about buying intentions, and these certainly give some indication of what's going on, but most of the time it's inconclusive. If the result is that anywhere between 20% and 45% say they are increasing their spending, 20-45% say they are decreasing their spending, and 20-40% say "No change in plans"—which is typically what you get from these surveys—then it's very difficult to claim that a new trend is afoot.

Another source of IT spending data comes from government statistics and Gartner/IDC trackers, but these tend to arrive rather late on the scene.

No, if you want hard data on the market, the earliest reliable material tends to come from the quarterly vendor financial results.

Last night, Dell published its results. Profits were down 17%, but revenue was up 11%. Sadly Dell didn't publish a constant-currency growth figure, so that will have to be estimated. However Dell did say that the earnings squeeze was a result of technology spending slowdowns and its expansion into Europe and Asia.

But is Dell telling the truth? Some watchers clearly think so: the chief investment officer at Solaris Asset Management, said: "[Dell is] saying lower IT spending is spreading. That is evidence of a global slowdown in IT spending. This certainly isn't good news for tech overall."

Personally, I believe the guy is being too hasty in jumping to this conclusion. We need to see several more sets of results from other vendors before we can confirm the conclusion. There has been a tendency in previous IT recessions for customers to stick with trusted brands. But is Dell trusted today? Dell is still in recovery, so there are a number of factors influencing its financial results.

So it may be just Dell. We need to see the results of many more vendors. In this instance, you cannot see the world in a grain of sand.

Averages are so simple, aren't they?

As I drove down the M3 today, I noticed a speed limit sign saying '50', closely followed by another sign saying 'Average Speed Check'. But are averages—or more precisely, the correct methods for calculating them—widely understood?

For example, if I drive for twelve miles at 40 mph, then another twelve miles at 60 mph, will I be breaking the average speed limit? The answer is No, because my average speed is 48 mph. {The formula for average speed is total distance divided by total time. Total distance is 24 miles; total time is 0.3 hours (for the first bit) plus 0.2 hours (for the second bit), i.e. 0.5 hours.}

But how many people can do the calculation in their head while driving?

Not that I would advocate Jeremy Clarkson's idea for defying average speed limits—i.e. bomb up the motorway at a crazy speed, then make up the total time by having a meal at a service station.

Another area where averages aren't quite what they might seem is in currency conversion. If someone asks you what was the average dollar:euro rate in 2007, you need to be aware that one isn't the precise inverse of the other. Check it out on oanda.com:

• The average rate from dollars into euros in 2007 was 0.73082
  
• But the average rate from euros into dollars in 2007 was 1.37074, the inverse of which is 0.72953.
  

The difference isn't due to rounding error. You can demonstrate it to yourself with a simple example—suppose for the first six months of the year, the euro was worth two dollars, and that for the second six months, the euro was worth one dollar:

• Thus if I trade one euro in the first six months, I get two dollars in exchange; and then I get a single dollar for my second euro in the second six months. So I get a total of three dollars for my two euros: average rate is €1=$1.5.
  
• But if I go the other way, and trade one dollar in the first six months, I get €0.5 in exchange. for my second dollar, traded in the second six months, I get €1. Thus $2=€1.5, i.e. €1=$1.333.
  

So the average rate depends on how much you exchange. It's an important point for the intelligence analyst calculating revenues overseas. And the difference between the two averages, however small it may seem, enables the wily trader to pull a fast one over the unsuspecting customer.

How accurate is that estimate?

(All the examples quoted in this article are over five years old—to minimise the risks to confidentiality.)

Internally we publish many numbers—for example, forecasts of market sizes and estimates of competitor revenues. But in most cases, we don't say how accurate those numbers are likely to be. Arguably we should publish, with each estimate of a number, an upper and lower bound, which define an interval highly likely to contain the true value.

Here is a sample of the maximum errors observed during 2002-2003:

Market Forecast for the year (made just before its start)

• Product Category (e.g. 'External Disk')
  
  
  
  • At the Country level: +/- 27%
    
  • At the EMEA level: +/- 25%
  
  
  
  
• Product Sub-category (e.g. a Server price-band)
  
  
  
  • At the Country level: +/- 34%
    
  • At the EMEA level: +/- 31%

Instances:

• For the year 2002, our forecast in 2002 was that 'External DASD' in the UK would reach $1,226m. But this 2002 estimate was corrected to $895m in 2003.
  
• For the year 2002, our 2002 forecast was that 'External Disk' in EMEA would reach $7,419m. But this 2002 estimate was corrected to $5,551m in 2003.
  
• For the year 2002, the 2002 forecast of 'Servers $25K to $100K' in the UK was $305m. But this 2002 estimate was corrected to $410m in 2003.
  
• For the year 2002, our 2002 forecast of 'Thin Servers' in EMEA was $319m. But this estimate was corrected to $219m in 2003.

A vendor's revenues for a quarter in the recent past

• Product Category (e.g. 'Servers')
  
  
  
  • At the Country level: +/- 47%

Instance:

• For 2Q02, one analyst firm estimated that IBM's server revenues in the UK were $93m. But according to internal accounts, our revenues were $63m.

It's worth commending the company for changing its numbers when they are known to be incorrect. One analyst firm, on the other hand, has historically not changed figures published in its server tracker when shown to be wrong. Instead it will adjust future figures in order to make the running total correct. So if any of the 2Q08 numbers turn out to be wrong, for example, it will adjust the 3Q08 figures to cancel out the error.

I suppose this approach has the advantage that, with each new release, you only need add the new column of data to your spreadsheet—i.e. you never need worry about previous quarters changing—but it does make you worry about the true accuracy of any individual quarter's estimates.

Another acquisition pairing not predicted: Infosys to buy Axon

There's a certain amount of embarrassment that comes from not forecasting that HP might acquire EDS—this year's mega-deal in the IT industry. I feel a small twinge of personal embarrassment at not predicting yesterday's news story: that Infosys would buy Axon.

I'd been watching Axon for the past nine years—since they went public—perhaps because they're based at Egham, just a few miles down the road from our Bedfont office. Equally, I knew that Infosys, a major Indian outsourcing services provider, was flush with cash, and was looking for an acquisition in northern Europe in the $300-400m zone. It just never occurred to me to connect the two together.

For what it's worth, here are the notes I've gleaned from the Web today and from my archive of clippings:

Infosys has announced its intention to acquire SAP services specialist Axon for about $750m, a 19% premium over the share price. The board of Axon is unanimously recommending the deal.

The acquisition will create the largest SAP services provider of the major Indian offshore players, and should place Infosys somewhere inside of the top 10 suppliers of SAP consulting services in the USA. Key clients include Xerox, BP, Orange, TXU, Transport for London, as well as major sub-contracting engagements in the USA, such as the Home Depot.

Despite the economic slowdown, SAP-related services remains a high-growth area. Infosys claims their SAP practice is growing 65%+ CAGR; other Indians are seeing similar growth rates. It’s not new applications that are driving growth. Instead the market is being driven by consolidation of multiple SAP platforms, upgrades to SAP’s latest releases—customers get penalised by higher maintenance charges if they stay on old versions—and the relocation of application management offshore. In summary, it’s all about cost reduction. The Indians that are seeing their SAP practices grow fastest say it's because they have the lowest delivery cost.

Axon brings a raft of clients to Infosys where it needs them most: in Europe. Infosys gets about 27% of sales from Europe, which should pan out to around £600-700m this year. The extra £120m+ of EMEA revenues that Axon brings will come in very handy, albeit at around half of Infosys’ margins.

”Our stated policy is to move up the value chain, move into the consultancy space,” said the CFO of Infosys.

The chief executive of Infosys said yesterday: “A lot of Indian companies are looking at Western Europe. This is our first acquisition there, and right now our focus is to make sure that this process goes through before we talk about where we go next.”

Axon Background

Axon was set up by Mark Hunter in 1994 after he left SAP to create a team of ERP services specialists.

Back in 1999, describing itself as a 'UK SAP implementer', Axon was available to provide statements for the press on the state of the R/3 market.

In 2000, calling itself a 'leading UK ERP integrator', Axon was happy to put the boot into the troubled Baan with a comment or two about its general unsuitability for most customers.

In 2002, Axon acquired an Australian consultancy which contributed £7m to its annual revenues of £43m.

Early in 2003, Axon was announcing an 87% downturn in net profits the previous year, due to the SAP market slowdown.

In 2006, Axon acquired three vertically focussed SAP consultancies in the USA, which contributed £32m to the company's total revenues of £138m. But it also disposed of its Middle East operation. Axon had just £6m in cash—which could have put its survival at risk if it expanded too quickly—so it arranged a £40m loan facility with its bank. Ovum, now describing Axon as 'the most spectacular beneficiary of the current revival in demand for SAP services', outlined the secret of Axon's strategy:

1. Qualify deals early,
   
2. Stick to a handful of verticals (particularly local government in the UK),
   
3. Concentrate sales effort on a handful of big contracts,
   
4. Ride on the back of a large outsourcer (especially Capita) that needs a C&SI partner to win deals. (This may give a clue as to Infosys's interest in Axon.)
   

Early in 2007, Ovum was lauding Axon's offshore investments in Malaysia, where it had 300 staff. Axon split the roles of CEO and chairman. Its stated goal was to become one of the biggest SAP practices in the world. Founder and executive chairman Mark Hunter said he believed Axon now had 1% of the global SAP services market.

In the second half of 2007, Ovum was describing Axon with words of praise that it normally reserves for Accenture and Oracle: 'Axon's model continues to work wonders ... Axon has become the poster child for the focussed buy-and-build strategy'. The UK public sector had been particularly good for Axon, with a large programme at Birmingham City Council. Axon bought another firm in the Far East, this one with 150 SAP consultants in China, Singapore and Kuala Lumpur. Its CEO described Axon Group as 'the largest consulting firm in the world focussed on the $24bn SAP services market'. But chairman Mark Hunter left the firm, and according to analyst Richard Holway, this made it inevitable that Axon would get swallowed up in an acquisition sooner or later.

In January, Axon shares halved in price for no clear reason. In March, Axon reported revenues of £205m, organic growth of 29%, and operating profit margin of 17%. It admitted that there was little room for growth in the UK; hence its focus on the USA and what it calls the DACH region of Europe: Germany, Austria and Switzerland.

So how did I miss this one? Was it simply the mismatch between Infosys's supposed target price range and the eventual price bid for Axon? Or is it that there are just too many possible pairings of cash-rich acquirers and potential targets?

Sources:

• http://www.feedingthesapecosystem.com/2008/08/infosys-to-acquire-axon-potentially-top.html
  
• http://hotviews.blogspot.com/2008/08/infosys-to-acquire-axon-for-407m600p.html
  
• http://www.axonglobal.com/pages/about_us/news/infosysbid.asp
  
• http://www.ft.com/cms/s/0/7798b574-730f-11dd-983b-0000779fd18c.html
  
• http://business.timesonline.co.uk/tol/business/industry_sectors/technology/article4606620.ece
  

(Bank Holiday)

It's a public holiday here in the UK, so readers will be spared an entry today from this blogger.

Of Actual Currencies and Plan Rates

Yesterday I voiced some frustration at the infinity of questions that can be asked of a Market Intelligence professional, and suggested one way to filter out the questions that should be discarded: "Will you take the same decision whatever the answer?". I gave the example of asking for charts to be done in both Actual Dollars and Plan Dollars, and suggested these requests create extra work for no discernible additional value.

Today I'll try to attack the question of when to use Actual Dollars and when to use Plan Dollars.

For at least the past four years, the exchange rate between the pound and the plan dollar has been £1 = $1.55. This isn't an exchange rate that anyone outside IBM would recognise, but a constant currency rate used inside an international company has two immense advantages:

• Plan rates are better for sales target-setting. Sales representatives (SRs) and their managers shouldn't have to worry about fluctuating exchange rates when deciding how they are going to achieve their annual targets, because exchange rates are beyond their control. Customers in the UK pay IBM for its products in pounds, not dollars. So if SRs targets are set in dollars, it is far better if they are in plan dollars, not actual dollars.


• Plan rates enable fairer year-on-year comparisons. If you want to measure success from one year to the next, then it's far better to measure sales growth of a sales representative, division or subsidiary in constant currency than at actual currency exchange rates. The reason is the same: exchange rate movements are beyond their control, so they should be punished or rewarded for these movements.

There is little point in requiring people to carry two parallel sets of figures around in their heads—both plan dollar targets and actual dollar targets—so in order to calculate their market share, they will want to see the market size calculated in plan dollars.

However, the rest of the world operates in actual dollars:

• American IT companies publish their results in actual dollars.
  
• The values of publicly announced contracts are stated in actual dollars / euros / pounds.
  
• IDC, Gartner, Ovum publish their estimates of competitor revenues in each IT market in actual dollars.
  
• If you state plan dollar numbers to anyone outside of IBM, they won't recognise any of them.

A company which translated all of this data back into their own internal constant currency could justifiably be described as too inward-looking. Every time that you saw an externally produced estimate of the size of an IT market, you would have to translate it into plan dollars in order to check whether it looked reasonable.

So at some point, it's best to talk in actual dollars, to avoid becoming too unworldly. Our principle tends to be as follows:

• Market size charts tend to be given at Plan Rate.
  
• Charts of competitor revenues are given at Actual Dollar Rates.

There is, as you might expect, a conflict when we try to apply competitor revenues to a particular market.

And I haven't even tried to address the topic of the best measure to use when there is high price inflation...

Don't ask the question if you'll do the same whatever the answer.

schnauze /ˈʃnaʊz/ vb (tr). to spend time seeking out bits of information for which one has no practical use, and in so doing, waste the time of others.

Attached below is a picture of Zelda, our six-year-old Miniature Schnauzer. We love her dearly, but when on lead, she has a frustrating habit of holding up our walks with almost every step, to sniff around at another piece of pavement, wall or tree. Clearly she is seeking scraps of information—probably about who else has been to the same spot recently—but why does she need to know? It can't be an interest in other dogs, because she's generally not keen on other dogs.

Strangely, she tends not to sniff around as much when she is off lead, so perhaps she does it just to annoy me. So in honour of Zelda, I offer up the word schnauze, which I defined as above. It's not original—a quick Google reveals that several other Schnauzer owners are using the term in an eerily similar way. And the verb may well mean something in its original German.

Staff departments in large companies can lapse into a degree of schnauzing if they veer too far away from the company's core activities, which I summarize as developing stuff, making stuff, selling it, and servicing it. It's very easy to ask questions, but if you have no practical use for the answer—or if your actions will be the same, whatever the answer—then you shouldn't have asked the question in the first place.

It's easy to schnauze a market intelligence department. There are so many different ways of looking at the IT market that, despite us feeling we may well have the best grasp of what's going on of any observer of the IT industry, for most questions, we don't have the answer instantly ready to hand—some research is necessary. For example, market intelligence has to measure many things in at least two currencies: not just the actual US dollar that most of the industry has agreed upon as the standard rate of measure for financial activity, but also the plan dollar that is used internally to help each country's operation measure its sales success. So if market intelligence produces a graph in actual dollars, why not then ask them how the graph would look in plan dollars?

Tomorrow I'll give my views as to when is the right situation to use actual dollars and when to use plan dollars. Hopefully someone will reply, and I'll learn something, and we can begin to distinguish the action-oriented questions from the schnauzing.