WELL Benefits Occupants More Than LEED: Commentary on a Comparative Study of LEED vs WELL

A recent Scientific Reports study finds that occupants are more satisfied with WELL-certified buildings than LEED-certified buildings. By our estimate, this translates to between a 1.16% and 3.72% increase in productivity - a statistic that supports the additional complexity of WELL certification for positively impacting occupant experience. 

The study compared a series of WELL certified buildings to a series of LEED certified buildings using occupant satisfaction responses from the Center for the Built Environment’s occupant survey database. Per the authors, while previous studies have compared WELL or LEED certified buildings against non-certified buildings, their results aren’t consistent, the number buildings compared sometimes limited, and their methodologies not always rigorous. This Scientific Reports’ study is an attempt to improve on those previous studies.

For those of us in the consulting world, such studies delving into the various details of building certifications provide additional insights to consider as we’re helping clients evaluate the pros and cons of various certification systems. Understanding the various impacts on occupant satisfaction is an important factor to consider when choosing a certification system, as it can be an indicator of occupants’ health, productivity, performance, retention, etc. We'll dive into estimated performance impacts later, but first, let's look at some of the key findings. From the study:

“Overall building and workspace satisfaction was high in both WELL-certified buildings (94% and 87%) and LEED-certified (73% and 71%). We found that there is a 39% higher probability of finding occupants who are more satisfied in WELL-certified buildings [bold added] compared to LEED-certified buildings, indicating occupant satisfaction is higher in buildings with WELL certification.”

In addition to the overall building and workplace satisfaction survey responses, fourteen IEQ, spatial, and functionality related satisfaction responses were also compared between the two sets of buildings (listed in Table 1 below). Figure 2 from the study shows the gap between the satisfaction response means for each of these other fourteen parameters examined from the CBE survey. For all of these parameters, the mean satisfaction response was greater for the WELL certified buildings versus the LEED certified buildings.

And except for “Amount of Space”, these differences in the means for each of the other parameters were statistically significant with non-negligible effect sizes. I should note that while non-negligible, the effect size values were still classified as small, indicating a meaningful, but not necessarily substantial, difference between the two groups (and that other factors besides the certification systems were likely impacting the satisfaction responses - see the limitations discussion below).

Given WELL’s larger focus on occupant experience and overall health and wellness compared to LEED, finding a greater impact on occupant satisfaction comparing WELL to LEED buildings shouldn’t be that surprising. But while the LEED buildings had overall satisfied responses for the building and workspace above 70% (very satisfied + satisfied + slightly satisfied), this is still under 80%, which is arguably not great. For these LEED buildings, almost 30% of the occupants were neither satisfied nor dissatisfied, slightly dissatisfied, dissatisfied, or very dissatisfied.

However, the study has limitations, many of which the authors acknowledge. We had some internal conversations regarding the limitations as well, and for us, the following are noteworthy.

Limitations

• Three of the limitations we observed expand on the authors’ first limitation: “While we masked the potential influence of several parameters on the outcome using the statistical matching approach, there remain other architectural features within the buildings or personal factors that affect occupant satisfaction.” These masked parameters consisted of certification, award level, years in building, time at workspace, type of workspace, and near a window.
• Unaccounted for correlations may be influencing the size of the gap in satisfaction responses between the WELL and LEED buildings in this dataset, potentially significantly (and likely contributing to the lower effect sizes). We agree with the authors that the age of the buildings and job satisfaction are two such unaccounted for parameters. Other parameters that likely also need accounting for include employee engagement, workplace attachment, degrees of personal control, and regional/cultural demographic factor variability.
• Unaccounted for correlations may also be built into the likelihood that organizations who choose to invest in WELL are also willing and able to more heavily invest in their employees overall compared to organizations only pursuing LEED. While this likely impacts the results, it’s also true that much of this “extra investment” by organizations is already built into WELL to begin with (reducing the potential bias).
• While the study may be lacking important control variables, the authors also noted that “… including more matched variables may unintentionally increase the variability of other matched variables. It is not known which control variables are most important for this comparison.” Here’s where other exploratory statistical methods, like factor analysis, could illuminated which control variables deserve greater focus.
• The authors rightly noted that “… occupant surveys are unable to characterize every aspect of the workspace experience or indoor environment”, limiting their ability “… to understand exactly why higher satisfaction was found in WELL-certified buildings.” This is where targeted ethnographic fieldwork or comprehensive POEs (that includes building assessments and IEQ measurements), conducted as part of the research, would provide valuable insights into the underlying why (as well as what control variables should be focused on).
• Additional insights would be obtained by adding uncertified buildings (as well as other types of certifications, like Fitwel) to the analysis.
• The WELL certified group of buildings only consisted of twenty buildings.
• Regardless of the limitations and the need for follow-up studies, there are still important implications here for the industry to take note of when applying these certification systems..

Implications

• The study provides some justification for the WELL Building standard’s degree of rigor, complexity, cost, and depth of underlying research (which sometimes acts as a barrier for using it). It would be worth seeing a similar comparison to Fitwel.
• These results are aligned with some of the criticisms levied against LEED v5 during the first round of public comments for decreasing LEED’s rigor associated with indoor environmental quality (IEQ) and occupant-centric metrics, except for indoor air quality (IAQ), which increased. This may increase the gap in occupant satisfaction between WELL and LEED (except for potentially air quality). Deeper ethnographic or comprehensive POE dives would provide critical insight into how this might play out. 
• This study provides another important data point to use with clients when helping them select the most impactful certification system for their circumstances. As illustrated below, its potency is further amplified by integrating quantified estimates of impacts on occupant performance. Are you ready to rethink your certification strategy? Let us know.

People-Cost / Benefit Comparison

Below are the results of an analysis estimating the difference in performance impacts of WELL vs LEED, looking at three of the IEQ related parameters: Air Quality, Visual Comfort, and Noise. To do this, we made use of BranchPattern’s happē™ tool, specifically the IAQ Satisfaction & Performance module, the glare portion of the Daylight Access + Glare module, and the Distracting Noises & Productivity module, as these three modules are set up to take advantage of satisfaction / rating responses. A more recently created fourth module was used to estimated the combined impacts of these three IEQ parameters.  

To conduct the analysis, the mean rating values were estimated from Figure 2 of the study using the reported mean differences to improve the estimates; these were then fed into the respective happē™ modules. The calculated performance deltas, based on these estimated differences in the mean ratings between the LEED and WELL buildings for air quality, visual comfort, and noise, are shown below in Table 2. Performance in this case generally refers to office and education related tasks (typing, reading, comprehension / learning, and math).

Most of the happē™ tool modules aren’t currently configured to provide a +/- range around the reported value. We considered using the reported 25th and 75th percentiles of the rating distributions as a proxy for the range in performance deltas for each of these. In the end we chose to simplify things, be conservative, and use happē™’s reported values as the Maximum Estimated Performance Delta and multiply these values by the effect size to obtain a proxy for the minimum value. These are shown in Table 2 above as well as Figure 1 below. As the effect size is an indication of the magnitude and practical impacts of the parameter differences between the LEED and WELL building populations, we thought it appropriate to use it for obtaining the proxy of the minimum value.

Table 3 and Figures 2 and 3 below provide an estimated dollar impact of these performance gains of WELL relative to LEED. These are based on a hypothetical organization occupying a corporate office environment with 1,000 employees having an average annual salary of $90,000. These estimated amounts essentially represent salary dollars spent more effectively relative to office type tasks.

However, the above estimates assume everyone is in the office 100 percent of the time. If everyone were only in the office 50% of the time, then the estimated average combined impacts would be:

• Per employee: $522 - $1674
• Per office: $522,000 - $1,674,000

One important note relative to IAQ. The performance and dollar amounts discussed above are based on satisfaction responses correlated with productivity / performance impacts. Humans can only sense a limited number of airborne toxins and pollutants, and the studies used for the happē™ IAQ module likely missed some potential contaminants. Therefore, these estimates likely understate the true impact.

Additionally, just as there are combined effects for IEQ impacts (and their associated satisfaction responses), so are there interactions among all responses to survey questions. The responses to these three IEQ conditions may be impacted to varying degrees by the other factors being asked about. This, and the fact that satisfaction levels or ratings only provide a “fuzzy” picture of actual performance and health impacts compared to direct measurements, it’s important to recognize that your mileage will vary.

The occupant satisfaction results from the original study, supplemented by the estimated performance impacts, provide developers, building owners, and design teams additional insights to consider when evaluating the pros and cons of various certification systems for their projects. Certification systems with varying areas of focus are bound to have different outcomes. It’s important that we keep working to understand those differing outcomes and their impacts on occupant experience and building performance. In the meantime, we can help you navigate what is known about the effectiveness of existing certification systems.