Management by Inflight Magazine

The press rang with claims that Marissa Mayer had ban from working at home a month ago (Feb 2013) citing her review of VPN logs as the basis for her decision. My issue here is not whether this is the right or wrong thing for Yahoo! but should others follow her lead and whether using the found data (i.e. data collected for other purposes) to support their decisions will lead to a good outcome.
Working from home is a complex task and challenging for both the worker and the company. I will look at 3 aspects of this decision:
  • Types of work
  • Types of workers
  • Data collected for other purposes

Types of work

A decade ago I worked as an academic and researcher (a creative inventive and innovative role) and in that role I would often take a day to work from home to write the first drafts of research papers. This work was inhibited by the presence of others, writing was not an easy task and distractions from it were welcome. Retreating from the work place was one way in which this task got effectively done. Currently in a technical role I use time away from the office to wade through dense technical documentation where concentration is essential.

My contention is that there are times and places for different types of work. Some aspects of a project benefit from interaction (innovation and creation) and others are diminished by interruption and communication overheads.

Types of workers

Some people and roles are not well suited to working from home. Self starting is critical and determination to complete tasks is essential. Requiring close supervision generally doesn’t go well with working from home.

Data collected for other purposes

Firewalls, web servers, VPN servers and routers tend to collect lots of data. Unfortunately, this data describes connections from machines to other machines. What it almost never describes is actually what a manager wants to know: what a person is doing e.g. how long a user reads a particular web page or what someone is actually working on. Unfortunately, it is not uncommon to use this data as a proxy for what the manager wants to know. Using data in this way is fraught with danger and should new supported by direct observation or surveys to ensure that the proxy is a reasonable approximation for the actual item that you wish to measure. In particular, it is extremely dangerous to use a proxy just because someone else has. You need to check its validity for your circumstances.

VPN logs would not be a good proxy for productive activity for the writing of papers or reading hardcopy documentation.


At issue here is not whether working form home is good, bad or appropriate for your firm; the key is to determine if it is suitable for you on a firm by firm basis (possibly a worker by worker basis) using appropriate measures rather than just following a management trend.