Most ‘enterprise’ applications are not actually used across the entire enterprise. That does not mean to say that there is no enterprise benefit arising from them but only a proportion of the workforce will use all the functionality of the finance system or the HR database. One notable exception (excluding the intranet) is an enterprise resource planning (ERP) application which may be quite widely used even though few employees will be using every element of the application. However search applications are widely used across the organisation and in my opinion are just as business-critical as an ERP application.

For many years now I’ve been beating the drum for the need to have an appropriate level of support for search applications but I still get strange looks from IT Managers (who usually have budget responsibility for search) when I start talking about the need for a search support team. I thought it would be interesting to see what typical support levels would be for an ERP application. Computer Economics, based in Irvine, California, has been reporting on ERP user/support ratios since 2012 and published its latest report in March. The median level of support turns out to be 58 users for each support team member. This figure is supported by a report from Janco Associates and from a number of other sources.

I am well aware that ERP applications have a very high level of complexity and frailty but even if this is an order of magnitude more challenging than search then the ratio for search would be around one member of support staff to 500 users. One further level of magnitude and you get to 1 per 5000 employees. You can do the maths as easily as I can but that says that a reasonably-sized multi-national company with 50,000 employees should have a team of ten supporting search.

It will be interesting to see what the outcomes are of the annual Findwise survey when they are released in September but my guess would be that the majority of organisations will still have just one person looking after the entire range of search applications. When I come across low levels of search satisfaction it is rarely a technology problem. The challenges lie with poor quality content, variable (if any) metadata and no resources to monitor and respond to user requirements.

Martin White