Part I: Next Generation Support

7 Cups is built on the belief that every single human deserves access to free emotional support Weve made good progress towards this goal over the last almost 3 years. We started doing 50 conversations a week and now we do 130,000. We had a small handful of listeners and now we have over 160,000. We were just US based and now we have representation in 189 countries providing support in 140 languages. Im really proud of our collective work and how far weve come.

Weve also worked closely with researchers to better understand how to help people. For example, our first study looked at the therapeutic alliance levels (the degree to which members feel that listeners care for them and are aligned with them in meeting their goals) met by listeners and also highlighted how people benefit from authentic support provided by listeners. Additionally, we worked with a researcher that did initial research on 7 Cups as an adjunct to treatment for people struggling with Schizophrenia and a separate study supporting those experiencing perinatal depression.

These studies have yielded a number of insights. As one example - though the sample size is small - we found out that those struggling with Schizophrenia tend to use 7 Cups primarily between 10 PM and 8 AM. This small insight has to be proven out, but if it holds, it has implications for providing better support. Most of the mental health world is asleep at this time, so we need to find ways to provide more comprehensive support throughout the evening hours.

Many of our most valuable insights were actually stumbled upon by accident, i.e. we were looking into one thing and discovered something else.This largely occurs because we are biased and limited in our thinking. The human brain comes pre-wired with limitations. For example, you can only see light within a certain spectrum or hear things at a certain frequency. Think of dog whistles as an example. Humans cannot hear them, but dogs can. What complicates things further is that our biases often trick us into seeing things not as they are, but as we imagine them to be. Another example - when the moon is low on the horizon - like just over the trees or the buildings - we say The moon is huge tonight! or Look at how big the moon is! The moon looks so much bigger when it is on the horizon because we compare it against the trees or the building near it. When the moon is high in the sky our eye has nothing to compare it against so it looks smaller. The moon, however, never really changes size. It just looks like it does.

These are sensory examples, but they illustrate how we bring all kinds of biases and assumptions to understanding our world. Research, when done well, helps to remove these biases and helps us better understand or map how things work. When we have accurate maps we can make more progress and help more people. When our maps are inaccurate - or not complete - we make little progress, or worse, we get lost. For example, if you are in Boston and I give you a map of New York, then you are going to be very frustrated. Why? Because you have the wrong map.

In the world of emotional support and mental health, we still have many outdated or, unfortunately, wrong maps. This isnt anyones fault. Nobody is deliberately trying to make bad maps. It is just a consequence of the limitations that we have as humans and groups of humans trying to make sense of complex processes. Fortunately, like with all other areas, technology is helping us make much faster progress in ways that were not possible before. Steve Jobs famously said that the computer is a bicycle for the mind. It is a way to extend our capabilities, insight, and analysis. Computers help us build better and more accurate maps.

One challenge that the broader mental health system has faced is that there is little local research done. Most clinicians do not provide a pre-therapy measure to assess the persons symptoms. Similarly, most do not provide a post-therapy measure to assess progress and give the person feedback on how far they have come. Further, none of this data is housed anywhere so that patterns can be seen or better understood. Instead, even when these instruments are given, they are just filed in a cabinet somewhere. So, unfortunately, minimal learning occurs on individual and group levels.

7 Cups is different. We designed it to be different deliberately. Early on in both chat and growth path steps you can take a wellness test to see how you are doing. At certain intervals, you can re-take the test again and again to see how much progress you have made. I know this seems really simple, but, believe it or not, it is transformative when compared with the more traditional model. We can actually track progress and explore what is most helpful to different groups of people. We have a study under review now that shows both African American men and Latino men benefit the most from 7 Cups. There is a lot of stigma against these two groups seeking help. Because 7 Cups is anonymous, people in these groups feel much more comfortable seeking help. Isnt that remarkable? We would have never learned that if we kept doing things in the traditional way.

We are a small team, so our ability to run analyses against all of our data to make 7 Cups better has limitations. Weve partnered with brilliant researchers (see our research page here) and weve made progress, but still our progress is limited. Fortunately, technology is now in a place where we can begin to create programs to do research for us. These programs can look for answers and help us make 7 Cups better. These programs do not have the same inherent biases that the human brain has, so they can uncover interesting insights that we would have otherwise missed. Research will help us to help more people, more effectively.

We are just in the very initial stages of conceptualizing how to do this next level of research. Like our other studies, itll involve working with world class researchers (like Qntfy - see below for details) with expertise in data science, machine learning, and statistics. Here are some of the initial studies and improvements we want to make:

Listener Matching - Right now we have a really crude member to listener matching system. The member indicates their issue and listeners either opt to help that person or not. This is good, but it isnt anywhere as good as it could be. With time, we may be able to look at patterns across messages to see what listeners are especially good at helping people with depression, anxiety, work stress, or PTSD. It is very likely that we have thousands of listeners that listen in a particular way that provides heightened support for people facing these challenges. At present, we just cannot see these listeners. We have no way of differentiating them and, of course, no way of matching members up with them.

Also, this would not limit people being able to take chats outside of proven areas as we all have room to grow and learn. It would also have no impact on people choosing or selecting listeners from browse listeners, group, forums etc.

Quality - Right now our quality hovers right around 80%, which is very good. Many people have worked very hard, and many initiatives undertaken, to help boost quality above 80%. Weve not been as successful as we would like...yet. We believe we can use these capabilities to significantly boost quality. Some therapy sites do the following with therapists. Imagine a scenario where a listener is talking with a member and we provide 3 potential options that the listener can select to respond. Listeners could respond as they like or they could opt to drop one of the recommended messages into the chat (the messages will be able to be edited).

New therapists often review transcripts of other therapists or watch videos of therapists doing therapy in order to better learn. You can see this as a more real-time example of this training process. Once the person felt comfortable as a new listener, they could turn these tips off (they could also choose to turn them off from the very beginning). The idea is they can help with quality and training until the person feels confident.

Over time, wed learn more about active listening and how certain listening styles help people overcome the challenges they face. Well share what we learn with others through peer reviewed journals, so that other communities can also benefit. In terms of a time line, I dont foresee us adding this capability before spring of 2017 and it might even be further out. This is going to take a while.

I want to also be very clear that our research and our own experience underlines how important our common humanity is to 7 Cups. So many healthcare settings are clean and almost sterile - like they work to remove the human side of things. We do not ever - and we will never - let that happen on 7 Cups. I often refer to this as the grittiness of who we are. This grittiness - lived experience - is authentic, real, and a big part of what makes us approachable. Technology - whether it be messaging, exercises, or these training tips - is always used to facilitate greater human connection. We do not need more technology, we need more personal connection. We will make sure that we do not ever compromise on the genuine and real human approach that makes 7 cups unique.

Zooming back out, what does this mean for 7 Cups? It means that well be able to build better and more accurate maps of the challenges we face.

What does this mean for you as an individual? Your experience on 7 Cups will not change, but as you go about giving and receiving help, we'll be able to use advanced programs to help us learn together what generally works, what generally doesnt work, and what works well for you as an individual. We'll gain an understanding about how to help people better. You will have the ability to donate your data or opt out of this program (it is still quite a ways out) in your settings area. If you decide to opt out then we will continue as we do now and not utilize any of your anonymous data to help us build out this map. If you decide to donate your data, then we will use all of your data - including all of your messaging data - to help us build out this map. Your information, as always, will be completely anonymous. The individual data is not the important part, the important part is the aggregate data. It is looking at the patterns of listeners that provide help to people that are struggling with depression and seeing how they tend to listen or say things that help. It is looking at how people that are anxious make that transition from being overwhelmed with worry to worrying less, to not worrying at all.

This is not going to be a silver bullet. We are not going to be able to immediately apply what we learn and help people across a variety of issues, but we are going to keep at it and we are going to keep learning. That is the goal. Imagine, no matter your issue, if you were able to come to 7 cups and get matched up with the best listeners, tailored to your issue and the way you see the world, to help you overcome the specific challenges you face. Wouldnt that be amazing?

This is still quite a ways out for us. I wanted to make this long post to share my thinking with you and answer any questions you may have. As always, thoughts and feedback welcome!

About Qntfy:

Note from Glen: I really like these guys because they are in this for the right reasons. The people on their team have been touched by suicide, trauma, and other issues. Like us, they want to make an impact here because they understand what it feels like to struggle.

Who we are:

Qntfy is a company made up of data scientists, computational linguists, psychologists, researchers and advocates for change in mental health. Qntfy is a firmly person centric company. We dont try to do this alone. We have a world class advisory panel that is unique (so far) in our industry. Unique, because our advisory board is made up of a carefully balanced mixture of computer scientists, researchers, practicing clinicians, activists with lived experience of mental illness and leaders in the suicide loss survivor community. We believe firmly in the "nothing about us without us" concept. No project related to mental health gets to the planning stages at Qntfy without input and approval from our panel and other outside expert opinions. We are dedicated to responsible, compassionate work and we know it affects the real lives of real people. We serve people by creating tools that people can use to help themselves and others. We have one goal: to empower people with information to guide their decisions about their well-being. We firmly believe that the old days of labels, prejudice, discrimination and stigma of mental illness are coming to an end and we want to play a part in seeing that happen sooner than later.

7 Cups

Our partnership's goals are simple, to provide better faster more intelligent care by dynamically improving the way the system matches listeners with visitors. With time, we hope to get the chance to build even more intelligent tools that will empower listeners, guests, and members to rapidly understand the state of their conversation. We take personal privacy very seriously and we will do all of this in the most privacy conscious manner possible. We dont do manual qualitative coding of each individual conversation so no person would be reading any identifiable chat information. The vast majority of the feedback and training (human feedback that makes our models more accurate over time) will come directly from 7 Cups listeners as part of their existing interaction with the site. Qntfy will only ever receive anonymized and de-identified data. We use industry standard security protocols (like those that already protect this data at 7 Cups) and we make no compromises in this area. We segment duties (Qntfy personnel only have access to the data they need to do their jobs). We use encryption liberally and all traffic to, from and between our servers is protected by industry standard methods. We understand the need for trust and anonymity when discussing issues like these and we hope to build the same level of trust with the 7 Cups Community as we have with our other partners and collaborators.

Here is a short (incomplete list) of the partners and collaborators we have worked with in the recent past:

The American Association of Suicidology
Johns Hopkins University
#SPSM (the online community dedicated to Suicide Prevention and Social Media)
Georgia Tech
Georgetown University
George Washington University
Texas Tech University
Stony Brook University
University of Colorado Boulder
University of Pennsylvania
The State of Massachusetts
The State of Montana Office of the Suicide Prevention Coordinator
The State of Texas Suicide Prevention Office
Mental Health of San Francisco