An IEA research Task on Optimization of Solar Energy Use in Large Buildings established that the most important first step in designing a successful low energy solar building is to establish what metrics might be used to eastablish success. For example, with lighting quality, the client, the energy analyst and the lighting analyst and the architect (who may be one and the same person!) must establish how they are to measure success in the context of the work or activity in the building. The whole design team should be involved in establishing and agreeing what the daylight and energy efficiency performance targets are and how the team are to measure (the ‘metrics’) whether they have been met. Daylight Metrics that could be useful:

• daylight autonomy
• glare
• thermal comfort needs (LINK)
• desirability of variation / constancy of illumination conditions (LINK)

user needs - agreeing a target lighting quality specification
In the special issue of Energy and Buildings collected papers summarising the work of this of this IEA research Task, Reinhart and Selkowitz describe how in 2002, a paper in Science (D.M. Berson, F.A. Dunn, M. Takao, Phototransduction by retinal ganglion
cells that set the circadian clock, Science 295 (2002) 1070–1073) identified a previously unknown direct connection between the eye and the circadian pacemaker within the brain that drives daily wake–sleep cycles. They suggest that "This emerging research field could establish new requirements for lighting buildings that are based on non-visual effects of light on humans. Such requirements might at times be higher than the minimum visual requirements that are currently stipulated in norms. Satisfying these new needs could become both an environmental burden, if met through electric lighting, or an opportunity, if satisfied through an increased use of daylight." They then introduce a paper by Webb and a paper by Galasiu and Veitch and intended to "frame the ongoing debate among designers, lighting manufacturers, and other stakeholders within the design community of how future lighting requirements could or should be accommodated."
user needs - glare criteria and calculation tools need improvement
One of the principal reasons for the overglazing of buildings is the lack of easy to use criteria and tools to identify and assess glare conditions. And, one of the fascinating aspects of this IEA Research Task collaboration has been to discover the “feet of clay” on which current glare formulae are founded. To paraphrase Eliyahu Ne'eman at the 2005 IEA Research Task workshop at LBNL “ …Hopkinson was not keen for the CIE to quote his window glare formula as he believed that there had not been enough work done to validate it..” As there has until recently been no alternative research work available, the rough correlation between window and glare that Hopkinson formulated has been used in computer software that reports glare.
In the special issue of Energy and Buildings a paper by Wienold and Christoffersen investigates "... the relationship between luminance distributions in over 70 subjects’ field of view and their self-reported perception of glare. The outcome of this study is a new glare prediction model called ‘Daylight Glare Probability’ that has been implemented into a RADIANCE-based evaluation tool." Clear, Inkarojrit, and Lee observed 43 subjects in a full scale test room equipped with switchable, electrochromic windows (visible transmittance range from 3 to 60%) and Venetian blinds. Reinhart and Selkowitz observe the degree to which people's behaviour affects net energy savings as they trade off glare control and daylight admittance. Minimising glare often means switching the smart glass to its lowest light transmittance, increasing electric lighting use. "New studies split the facade into an upper daylight window and a lower vision window, thus providing comfort and higher lighting energy savings." Sutter, Dumortier, and Fontoynont's work is apparently consistent with other work report that show that each person will operate a blind or shades consistently, but the variation in operation between people is very large. Remote controls seemed to increase use of blinds and brighter VDU screens made people more tolerant of higher dalight levels.
user needs - controls
No matter how well the building is daylit, if the goal is also to achieve energy savings then there need to be good electric lighting controls. In predicting energy savings from daylight one of the larger unknowns has been the dearth of information available on real switching behaviour in real buildings. Papers in the special issue of Energy and Buildings by Lindelöf and Morel and by Nicol, Wilson, and Chiancarella analyse the use of light switches. "In accordance with previous findings, occupants chose to work under widely ranging lighting conditions and used their lighting and blinds to ‘adapt’ to their environment." Lindelöf and Morel have developed a model complementary to Hunt's 1979 paper that can be used as a starting point for simulating user behaviour.
Reinhart and Selkowitz comment: The results from all of these human subject studies are encouraging. They reveal that while our understanding of office workers’ use of personal controls remains limited, findings from earlier studies were independently confirmed in different buildings and countries. This result is important since human subject studies are expensive. The results further reiterate the usefulness of international collaboration.
is 'daylight factor' enough when dynamic simulation is possible?
The daylight factor has long been criticised. As a metric of daylight performance it suffers from being referenced to a single overcast sky condition - not only not taking account of sunny days and therefore orientation of the building, but also often defined as some international general 'CIE sky' that bears no relation to the local climate.Reinhart and Selkowitz note: Dynamic daylight simulations, i.e. time-series of indoor daylight levels over a whole year, are becoming an attractive alternative to daylight factor calculations. The important remaining question is how does one interpret these time-series and boil them down into accessible performance metrics? "Nabil and Mardaljevic compare daylight factor simulations with two dynamic metrics, daylight autonomy and useful daylight index (UDI). UDI is an interesting new metric that provides a measure of both whether minimum illuminance levels are maintained in a space through daylight, as well as how often daylighting levels might be too high."