Will technology make professional site sitelectors redundant?
Technology is “disrupting” a growing number of industry sectors. Retail, finance, news media and taxis are just a few obvious examples of sectors in which technology is changing business models and displacing traditional market leaders.
At the same time, automation, robotics and artificial intelligence are expected to displace not just manual jobs but increasingly sophisticated service positions too. Truck drivers, legal researchers, tax preparers, accountants…the list of professions under threat of disappearing is growing longer by the day. In many cases, this threat is real and the switch from humans to machines is already happening.
It must be only a matter of time before technology disrupts the site selection industry too and eliminates the need for professional site selectors who make a living advising companies on selecting the best locations for new facilities. After all, site selection is an intuitive and logical process, based on collecting standardized data and using relatively simple modelling to rank a manageable number of locations. Surely collecting and analyzing this data is something that a machine could do more effectively (and cheaply) than a person? If artificial intelligence can be used to fly aircraft and diagnose diseases, applying it to identifying the best location and site for a new manufacturing plant should be easy.
Technology has already changed the way site selectors work. The initial phase of the location analysis process, where relatively large numbers of locations are “screened” using basic criteria, has been automated and accelerated using relatively standardized and easily accessible data. Other parts of the process, such as criteria weighting or evaluating data using decision-making systems has also been automated using simple IT tools. GIS systems now provide a wealth of data and mapping opportunities that did not exist a decade ago.
The use of big data and automation is most advanced in site selection for retail, where micro-level data helps pinpoint the right spot for a particular brand or concept. The real estate industry uses similar data to determine which neighbourhoods are becoming “hot” and primed for gentrification. Sentiment analysis garnered from social media data provides further insights, for example for assessing how certain segments of the workforce feel about different cities.
While these changes are happening quickly, the overall site selection process – especially for large and complex corporate facilities such as manufacturing plants – has not changed all that much and is still performed by people. What is holding up radical disruption in site selection?
The limiting factor is the availability of data. To generate results, the data used for site selection needs to be accurate, up-to-date and perfectly comparable across many locations. In the US, the availability of data has improved dramatically, with recent and standardized statistics available across states, MSAs and counties. Economic development agencies in the US have contributed to this by providing a wealth of data, often tailored to the needs of site selectors. This explains why the use of technology and automation in site selection is far more advanced in the US than anywhere else.
In other parts of the world, data is nowhere near as available as in the US. Even in the European Union, statistics provided centrally by Eurostat are usually incomplete, out of date or simply not relevant to location analysis. Data collected by individual countries is typically too aggregated and not comparable across countries. Try finding out what an experienced metal fabrication assembly worker earns in Sibiu, Romania or how many German-speaking accounts-payable specialists work in Sofia, Bulgaria and the limitations of the available data become very clear. The shortage of data is even worse in some emerging markets where official statistics sometimes cannot be trusted or where even basic data simply does not exist.
The often-held belief that data is readily available – online and at no cost – does not hold true in most parts of the world. Even if perfect data were available for a specific location, this would not be very useful for site selectors who require comparable data for all locations being assessed. In many cases, this data needs to be found through old-fashioned primary research (i.e. speaking to other people). One of the attributes of a skilled site selector is the ability to obtain reliable data in places where no published or official data exists.
Another challenge to fully automating the site selection process is that every project is different. Each company has distinct requirements for workforce, suppliers, logistics and other criteria. This means that much of the data needed to select a site for each project is unique, especially for more complex projects involving a multitude of variables. Although there may be some overlap, the data used to pick a site for a food manufacturing plant is not going to help find the best site for a new automotive plant or shared services centre.
Since data for site selection needs to reflect current conditions, it has a short shelf-life and becomes obsolete very quickly. In the absence of reliable central data sources, this makes the task of updating the data both time-consuming and “manual”.
Site selection is also not a purely quantitative process. Once data is collected, it needs to be evaluated, which requires the ability to spot trends and uncover risks by asking the right questions. Gut-feeling based on experience plays a role too and is sometimes the decisive factor in making the final location decision.
The technology to “disrupt” the site selection already exists and is improving rapidly. Artificial intelligence is advanced enough that machines should be able to learn how to do what site selectors do. What’s still missing – particularly outside of the US – is the data to power digital site selection tools and allow them to make the right decisions. Perhaps someone is already working on a system that will disrupt the site selection industry and make human site selectors redundant. For the time being though, it looks like there is a still a need for experienced site selection professionals and that technology represents an opportunity more than a threat.
By Andreas Dressler, ©Location Decisions