Jump to Navigation

The official truth about the software industry

In my previous post, I looked at what the firm level growth figures look like for the last year. It is now time to adopt a higher level perspective and look at how the Finnish software industry is faring on the industry level. Toward this end, I retrieved a few data sets from the databases of Statistics Finland.

In the official statistics, most of the business that we include in the survey fall under the NACE industry code 62 “Computer programming, consultancy and related activities”. Although some companies, particularly some game producers, are registered as software publishers under the publishing industries (NACE 58), these are a minority both in the software industry and in that industry code. Because of this, a majority of publications that discuss the software industry focus on the information processing industries (62) only. The official statistics become available with some delay, and the latest numbers available now are for the year 2010. The figures go back to year 2007, when the industry classification system was changed.

The IT services industry and its subindustries

The first table that is presented below shows the aggregate statistics on the industry code level. The higher level code, 62, is broken into four subcodes. The 6201 is the closest match to software industry, but many software companies are also registered under other codes and not all companies under this code do software as their primary business. Nevertheless, the 6201 industry code is the best proxy that the official statistics provide.

On a higher level, the overall IT services sector (62) clearly experienced a boom in 2008, a bust in 2009 and stagnation or slight recovery in 2010. What is noteworthy that these figures are not directly affected by Nokia, because it is classified under the electronics industry. However, the challenges faced by the company are inevitably reflected in the revenue of its subcontractors.

Looking at the computer programming activities (6201), we can see that not only both the revenue and personnel employed decreased in 2009, but they have continued to decrease in 2010. There is a stark contrast with the consulting services (6202) that have steadily increased their revenue through 2007-2010. Also, pure consulting companies tend to be more productive, i.e. one person working in this industry generates about 15-20% more revenue than a person working in the software industry. This difference is puzzling considering that software product business is often characterized as high margins activity. Clearly, this is one issue that might be worth further analysis in our final report.

Companies Personnel Revenue (Me) Average salary (ke) Average revenue (Me) Revenue (ke)/personnel
62 Computer programming, consultancy and related activities
2007 4 534 39 142 5 659 47,8 1,25 145
2008 4 783 40 928 6 094 50,6 1,27 149
2009 4 857 39 797 5 682 50,9 1,17 143
2010 4 981 39 955 5 697 52,0 1,14 144
6201 Computer programming activities
2007 2 807 24 113 3 220 47,4 1,15 134
2008 3 008 27 122 3 927 50,5 1,31 145
2009 3 071 25 211 3 280 49,8 1,07 130
2010 3 110 24 862 3 258 51,4 1,05 131
6202 Computer consultancy activities
2007 1 454 5 304 788 45,4 0,54 149
2008 1 489 5 614 875 48,7 0,59 156
2009 1 480 6 262 967 51,2 0,65 155
2010 1 534 6 584 997 51,8 0,65 151
6203 Computer facilities management activities
2007 227 9 448 1 625 50,4 7,16 172
2008 229 7 889 1 261 52,4 5,51 160
2009 237 8 093 1 412 54,6 5,96 174
2010 243 8 287 1 414 54,1 5,82 170
6209 Other information technology and computer service activities
2007 46 277 26 35,4 0,56 94
2008 57 302 30 39,3 0,53 100
2009 69 232 21 33,6 0,31 92
2010 94 222 27 40,3 0,28 122

Startup activity - company founding rates

Many people close to the Finnish startup scene like to argue that we are now experiencing a startup boom in Finland. This is a view that is portrayed by some government agencies as well. The second dataset that I retrieved from the Statistics Finland addresses this trend: To which extend is there a boom in founding new companies in the IT sector (62). The number of companies between this table and the previous is probably because this list includes all registered companies, but the previous only those that have actually had revenue. It is common that a discontinued company, or a company that actually never was started, is left in the desk drawer to wait for better days.

Without further introduction, the figures tell a clear story: If there is an entrepreneurship boom in the Finnish IT sector, it is not visible in the official statistics. In fact, the entry rate (number of new firms founded as percentage of total population) decreased between 2008 and 2009 and continued a slight decrease in 2010. This begs the question: are the people who argue that there is a startup boom wrong? We doubt that this was the case: First, the startup boom was not on the radar in 2010, but – if it exists – is a more recent phenomenon. Second, these figures include all registration events. Considering that many new companies are just subsidiaries of existing organizations, one-man operations, or e.g. ways to officially recognize revenue from projects done as side jobs, real startups are in any case a minority in the population. Because they are a minority, fluctuations in registrations for other types of businesses would mask any trends in the startups.

Year Firms started Starting rate Closed down Exit rate Number of companies Change
2007 929 13.4% 592 8.6% 6909
2008 1050 14.3% 753 10.3% 7345 6.3%
2009 969 12.8% 664 8.8% 7559 2.9%
2010 1009 12.8% 515 6.5% 7903 4.6%
2011 993 . .

Dominated by large service companies

The final table illustrates the importance of large firms for the IT industry. The table below shows the aggregate figures of the three largest firms, ten largest firms, firms outside the top ten, and all firms. The table shows that the largest IT services companies (Tieto, Logica, Fujitsu) account for roughly a fourth of the revenue for the entire sector, and including the rests of the largest 10 companies raises this to one third. The table tells also several interesting details. First, larger firms seem to be more productive than smaller firms. While this is seemingly the case, it can be a premature conclusion: The category other includes also all part-time companies, which naturally generate less revenue per employed person. Although we believe that there is a trend in that larger firms are more productive, it is not nearly as strong as portrayed by the table below. Second, interestingly, the larger companies seem to have faced decreasing revenue, but increasing number of personnel between 2009 and 2010. The figures do not reveal if this is a universal effect, or just caused by one of the three largest. The revenue trend in the other firms seems to be the opposite and firms have slightly increased their business in 2010.

Companies Personnel Revenue (Me) Personnel/firms Revenue (Me)/firms Revenue (ke)/personnel
3 largest
2009 3 8 003 1 446 2 668 482 181
2010 3 9 173 1 395 3 058 465 152
10 largest
2009 10 12 886 2 210 1 289 221 172
2010 10 13 230 2 126 1 323 213 161
Other
2009 4 795 24 778 3 171 5.2 0.661 128
2010 4 915 24 932 3 256 5.1 0.662 131
Total
2009 4 805 37 664 5 382 - - -
2010 4 925 38 161 5 382 - - -

Conclusions

So if one was to conclude this blog post with a grand conclusion, what might be said about these three tables? The best characterization might be “Business as usual”: After the recession years, there are no clear trends toward better or worse. However, this is with the caveat that the figures are now a year and a half old, and much has happened since that. For example, Nokia’s problems started to manifest in layoffs only after 2010. Clearly, there are much that can still be said about our more timely survey data