The UK SIC codes were last updated in 2007, more than a decade ago. However, new markets emerge every day and the SIC code system often fails to provide accurate classifications for new markets and is still skewed towards the manufacturing industry. In an economy which is predominantly service-led, this poses a problem for new businesses looking to classify their activities.
Your ability to find and target the right industries using the SIC code system is also severely limited. SIC categories range from the very niche, “striking of coins” to the outright generic, “other professional, scientific and technical activities”. New sectors are being created every other week and today it is simply impossible to completely categorise every sector in a four or five-digit code.
Today’s search engines and voice-controlled operating systems such as Google and Alexa, have culturally trained us to search intuitively in natural language. Instagram and other social networks have introduced hashtags to help us categorise content and search for themes. The days of command prompt are long gone, yet searching under SIC codes is akin to using Microsoft DOS commands.
A company seeking to discover new customers or suppliers in the sector of low-powered wide area network internet of things, or “LPWAN”, which is a rapidly growing multi-billion-dollar industry, would find it virtually impossible to find a suitable SIC code. Yet imagine if companies were tagged with #LPWAN like in Instagram or Twitter – it would make discovery of such companies so much easier. SIC codes once helped statisticians understand the economy by categorising industries, but we should accept that the very fibre of businesses is changing at a pace faster than the creators of SIC codes ever anticipated.
There are also practical implications to forcing businesses to identify themselves with a four or five-digit code. One would argue that in a business-to-business context, it’s possible to find new customers or suppliers through the internet, or through expensive marketing lists. Yet it is the next step of taking this shortlist and conducting due diligence which complicates the process.
Typically, when a bank or company conducts its due diligence on a prospective customer or supplier, it would pull information from other sources. This might include information from Companies House, credit rating agencies and more. Therefore, if the query is to find all independent LPWAN device manufacturers with a turnover of between £500K - £1m, with a credit rating that is average or above, one can quickly see the limits of going from SIC codes to a comprehensive list of companies and their associated information using traditional methods.
SIC code system gaps
Take RESPONSE PHYSIOTHERAPY LIMITED for example. They are listed at Companies House with the SIC code 93130. But when we look at their website, we find that they wouldn't really be a fit for a gym equipment manufacturer targeting the businesses you'd expect under "Fitness facilities". Conversely, you can find a company like BREAKSTONE FITNESS LTD who would be a great fit but are listed as “General secondary education” with the SIC code 85310. By targeting the first code, we'd end up with some companies that absolutely don't fit, and we'd miss some that do.
There's another big problem, too. Even with 731 potential classifications for businesses, there are still entire industries which don't have a SIC code specific enough for them. To be honest, our first example of Response Physiotherapy is one that falls into this category. You'd expect that there would be a classification for physiotherapy in some form, but actually the closest you can get is 86900 – Other human health activities. Much better than their current classification. But if you're specifically targeting physios, then you're going to find plenty of other businesses in that group that you'll need to filter out.
Why SIC codes don’t always work
The first reason why companies can end up with a SIC code that doesn't actually describe their business is the one we have just discussed. When there's no relevant SIC code to put down, a company can't do anything but picking the closest available category.
The second reason that I know of is laziness. Some of you reading this will have registered a company, and part of that process includes defining which SIC codes your company operates under. It's easy to just get it done by picking something that doesn’t look too odd and moving on to the more important tasks you have when setting up a business. That's probably how a huge amount of Taxi companies that should be filed under “49320 – Taxi operation” end up with the rental companies in “77110 – Renting and leasing of cars and light motor vehicles” for instance.
The third reason is very similar to the second and is a combination of laziness or ignorance by a third party registering a company on their behalf. Often companies are registered by an accountant or solicitor involved with the inception of the business and these people are far less likely to care about whether it's done absolutely correctly. Again, the mindset “get it done, get on with something else” prevails, which is totally understandable.
The fourth and final reason I can think of is as a consequence of a business pivot. I can absolutely see why a company would forget to update its classification when going through a massive business change, although the argument could be made that most pivots, which result in a large enough change to be in the wrong SIC code, would probably warrant a new company anyway...
Addressing the limitations
The good news is that with technology and more business-related open data sources, a more accurate business category query and robust diligence are now possible. Through due diligence platforms, companies are “tagged” with a more accurate description of business activities. Knowing what is actually going on within the economy and within specific businesses opens up a host of exciting possibilities.
For policy makers, there is a potential to observe and more accurately measure the impact of the policies they make. Banks would be able to identify companies to lend to that previously were under the radar, and insurance companies would finally be in a position to underwrite risk on a company-by-company basis rather than by crude industry groupings. But the most important benefit is to businesses themselves.
In my opinion and experience, we don't need to throw SIC code targeting away completely, but we need to overlay and augment it with other sources of data. For instance, you could find a way to gather all of the businesses from somewhere like Yell.com (with their business classifications) and combine that with your company data to find the outliers in a given SIC code. You could theoretically run the homepage of every potential website you want to target through something like Google's Natural Language API, to check whether the text on the page closely matches the subject you'd expect that company to be talking about.
By looking beyond the limitation of SIC codes and embracing the power of tags, businesses can more easily discover and build bridges with new trading partners, both in the UK and beyond. For those who continue to seek growth, the time to do this is now.