Redesigning Mail for San Diego's Housing-Insecure

We narrowed our scope to a specific day center that had trouble with distributing mail to its patrons in a timely manner. This inefficiency stemmed from many problems — most importantly its understaffing. The center could have as many as 150 patrons visit in a single day, and with only two or three employees present at a given time there were many instances of the demands of the community far outweighing the capabilities of its service members.

This led to unrest among the patrons — who believed the center’s employees were unfit to aid them — as well as unnecessary stress for the employees who felt the pressure to serve their community. Due to complications often associated with homelessness — such as using an alias to protect oneself against an abusive partner or having several pieces of mail for social benefits arrive for one person at the same time — finding all of one patron’s postage could take as long as 10 minutes. When the entire time period for mail distribution is only 30 minutes long, the volume of patrons and mail quickly becomes too overwhelming for the staff.

Their system consisted of a large cabinet set, organized alphabetically, and a line-up system for patrons. Each visitor would line up, present their name to the staff member, and wait for the staff to find their name among the mail. Once the mail was delivered to the patron, the next patron would present their name. This would continue until either the line ended or the 30 minute period ran out.

From our research and interviewing, we came up with several personae to represent the users — patrons and employees alike — who were involved at the center. Stakeholders ranged from those directly involved — the specific staff members and patrons — to the surrounding community (such as the government officials overseeing the budget for public aid and the citizens of San Diego). For the sake of our project, we focused specifically on the primary stakeholders: patrons and employees.

We had several factors hindering our progress throughout the project. First, we had our time restraints. As this was a school project for a class spanning 11 weeks, we only had about 9 weeks of usable work time to conduct all of our observations, analysis, and iteration. In addition, the day center we chose to work with was far removed from our campus and our access to transportation was limited. Between these restrictions and the limitations of 6 college students with other obligations, we found it difficult to make physical trips to visit our stakeholders.

Our other primary difficulty was access. As outsiders who wanted to reduce our disruption on the center, we did not have unlimited time to spend interviewing staff and patrons. We also did not have very much influence over changing the physical infrastructure of the center. They had problems with ramp accessibility, but we had no reliable way of creating a prototype that could be implemented in the actual center.

However, we were able to navigate these challenges by choosing a scope that we could influence — while we couldn’t purchase new cabinets for their mail system, we could restructure the methods by which they distribute mail.

We also made sure to visit the center in small groups — 2 or 3 of our 6 members would visit at a time so as not to draw attention to ourselves and create as little disruption as possible. We met with staff during breaks, rather than taking their time during operational hours. We talked only with patrons who were willing to chat with us rather than bombarding a random sample of patrons who may have been overwhelmed by our invasion of privacy. These measures may have introduced some research bias — as the qualitative data we collected would be inherently skewed by the types of personalities willing to talk to us — but we decided it would be worth the cost in exchange for reducing our interference with the day center.

Seeing as we could not remodel the center’s physical space and could not provide the center with more funds to hire extra staff, we instead designed a distribution system. This was more than a single organizational system — it was a complex of behavioral procedures, physical organization, and paper artifacts.

The most significant change we made to their system was a signup sheet. Originally, a line would form through the 30 minute distribution period. This was not only inconvenient, but an actual fire hazard, as it blocked a main pathway toward the exit. Each patron had to stay in line to reserve their spot, which took time away from them receiving other services like showers. By introducing a signup sheet, the patrons would reserve a spot in line without needing to remain there physically. They would only need to be within earshot for hearing their name called. This would reduce physical strain as well as impatience, as patrons could now wait in the comfy “living room” area with couches and sofas.

The second change would be the “signup period,” which would be held for 15 minutes BEFORE the distribution begins. This time is dedicated for the patrons to be able to begin the process while the staff are completing the mail pre-sorting. Pre-sorting was part of their original system, but combined with the signup sheets it becomes much faster to find a specific piece of mail for a specific name.

The signup sheets would be sorted alphabetically, as it reflects the current physical organization of the space. Each sheet would correspond to its own small section of the alphabet — A-B, C-D, etc. This forces the patrons to sort themselves, but this distributes the cognitive work in a way that offsets the load from the staff. As there are only 2 staff members operating on the mail but several times more patrons, it becomes easier for the large group to sort themselves than for there to be a “cognitive bottleneck” when it gets to the staff’s turn to find a name.

The progress of the line could screech to a halt because of a single patron, but if the signup sheets are stratified alphabetically, the staff only needs to find mail out of one drawer for that group, rather than needing to search the A-B drawer for “Abrams, Mary,” switch to the S-T-U drawer for “Smith, Jane,” then switch again to M-N-O to find “Miller, Sarah.” By stratifying the sign-up sheets, we reducing the amount of thinking — and therefore the amount of time — required to find each name AND transition from one name to the next.

Because of our promise to not interrupt the center’s work, we were unable to test within the facility. This was a significant setback, as we couldn’t see how our system works in realistic conditions.

To get around this, we conducted a roleplaying exercise with 5 of our 6 members. Two members acted as the staff — reflecting the collaborative role that two staff members would have in distributing mail. Two other members roleplayed as patrons, who would wait to hear their names called and retrieve their mail when called. The fifth member recorded the exercise — which you can view below.

This has its own limitations of not being faithful to the volume of mail and stressful conditions of the day center, but it provided useful insight about how the sorting would work and how much more space the front desk would gain without a long line in front of it. It was also rather fun to read the fake names from our fake mail — handcrafted and sorted for each round of the experiment.

Our testing yielded several insights — the biggest of which was that our sorting process was slowed down by our labeling conventions. Up to that point, we had been naming categories as “A-C,” “D-F,” and so on. But while we initially thought these categories were clear, we quickly discovered that if E fell between D and F, it took us just a little longer to find the right category to sort mail into. It only took us a slight bit longer, but the slight bit added up while sorting only 20 pieces of mail. With the added volume of the center’s 150+ patrons, combined with the added stress of sorting in a short timeframe, this cognitive delay was unacceptable.

The fix for this problem was simple and had a significant benefit — labeling the categories as “ABC” instead of “A-C.” It seemed like common sense after we made the discovery, but we were surprised that such a tiny detail meant so much to our project.

In the end, our final product was as follows:

During the 15 minutes before mail distribution, the center’s staff puts out the alphabetical signup sheets and begins sorting the day’s incoming mail into the cabinets. While patrons put their names on the sheets, the ensures the mail is properly sorted. Then when the distribution session begins, one staff member calls out the first name on the first signup sheet while the other grabs the mail from the alphabetical cabinets. Once the patron comes to the counter and retrieves the mail, the next name is called until the list is finished. All incoming mail that was not claimed remains properly sorted, in case the recipient comes another day and puts their name on the list.

(Pictured: Our team with our professor Taylor Jackson Scott, third from the left. Our final day was a presentation, and I was delighted to have completed my last class of my college career!)

This project meant a lot to me because of its topic — we were helping a struggling organization and people with housing insecurity. But it was also a personal milestone: an exercise in leaving my comfort zone to learn about a topic that required humility, self-reflection, and compassion. This project was not about university students learning about homelessness, it was about impacting the world in a positive way.

If I could do anything differently, I would definitely conduct my research within bounds that weren’t so restricting — because this project was ultimately for a 10-week course and two of those weeks were spent deciding a topic, we did not have nearly enough time to tackle some of the hardest problems with the project. We could not test within the center with actual patrons, and could not introduce any hardware changes (such as a new physical layout) due to financial restrictions. We were six students, and the center was already too busy for renovations. But despite that, we were still able to provide a helpful solution to an important problem.