Anatomy of an Android app launch: The CBT Keeper experience

In June of 2013, I was finishing a medical research fellowship when I came up with the idea behind CBT Keeper. CBT Keeper was imagined as a simple Android app that would allow users to implement the basic principles of cognitive-behavioral therapy to help them fight back against depression and anxiety. Although I had almost no experience with programming, I decided that I would learn to design and build the app myself. I believed (and still believe) that the world needs more high-value, low-cost mental health resources, and learning how to build this app has been a great introduction into the business of software. I can’t predict exactly what my future projects will be, but I am enthusiastic about the potential that software products have to empower mental health consumers and professionals.

Now that CBT Keeper is on the market, I thought I’d share a few thoughts about the progress I’ve made. This post is more for me than for anyone else, but I thought it would be valuable to lay out some raw numbers. Data can be empowering or paralyzing, but I hope that this data gives me a baseline upon which I can build in the future.

Bare Metrics:

Estimated hours spent building the app: 400 (including learning how to code and all of the associated marketing so far)

Number of weeks available in beta: 23

Downloads in beta: 1595

Number of weeks since full-market release: 2

Total downloads (including beta): ~8100

Revenue from in-app purchases: ~$450 (30% of that goes to Google or Amazon)

In-app purchase conversion rate: 2.2%

Average revenue per in-app purchase: $2.54

Revenue from advertising: $70

Estimated Gross profit: $385

Approximate gross profit per download: $0.048 (many in-app purchases come later, so it’s likely a little bit higher).

Things that surprised me:

Being an amateur developer is incredibly fun. Once you realize that the only thing holding you back is the time to read and try things out, it’s amazing what you can build.

Marketing is much, much harder than making an app. This shouldn’t have surprised me, but publishing one app has given me a ton of ideas about how I will incorporate marketing elements into the next one from day one of development.

A single news source can do wonders for a free app. I pitched a ton of sites and only had one response, however, that single article drove almost 5000 downloads. I’ll be leaning heavily on this fact to drive future campaigns.

If you have a mission, people will be happy to pay for your app. My mission is to make CBT fun, easy to use, and available to everyone. I had always heard that conversion rates from free to paid are approximately 0.5-1.0% on Android, but CBT Keeper has consistently driven rates over 2%. Not only that, but the app has a “pay what you want feature,” allowing users to spend $1.99, $4.99, or $9.99 to remove ads and password protect the journal entries users make. It has been delightful to see how many people chip in more to help me fulfill my mission.

For developers only: Don’t even think about using android.app.Fragment. This seems to be implemented poorly by several Android OEMs and produces errors that everyone but me can reproduce. Instead, you MUST use  android.support.v4.app.Fragment. This error alone dropped the app from 4.8 to 3.5 stars at launch. We can work with a 3.5, but I’d much rather have the credibility of a higher ranking going forward.

Future directions:

1. Polish CBT Keeper on Android and incorporate some of the key feedback given by users. There’s a lot of feedback I will have to ignore, but many important points do keep bubbling to the surface again and again.

2. Expand marketing efforts with giveaways on major Android blogs. Paid acquisition on a per-user basis is completely unfeasible, but I want to see what kind of ROI we can drive with a more mass-market approach.

3. As more revenue comes in, hire a developer to build the app for iOS. I had thought about doing this on my own, but I’d rather focus my efforts on learning from users rather than learning a new framework. Plus, having the app on both major platforms will provide excellent opportunities to promote it on sites dedicated to mental health consumers and professionals.

4. Reinvest proceeds into more great mental health tools! I still have a little less than 5 years of training before I can really devote a substantial portion of my time to Euthymic Labs, but I want to keep learning and developing things that make an impact.

Thanks for reading! If you’d like to read more posts like this, please sign up for my biweekly mailing list!


Depression is not a chemical imbalance

Depression is not a chemical imbalance–that’s what the best data on the subject suggests.

Now please don’t misunderstand me. I believe that major depressive disorder is inherently biological. I just remain unconvinced that brain chemistry alone tells us much about depression. At it’s best, the “chemical imbalance hypothesis of depression” is an outdated theory. More cynical observers might say that it’s an oversold marketing message.

The belief that depression is a chemical imbalance comes from a misunderstanding of the monoamine theory of depression. According to this theory, depression is caused by dysfunction in one of the monoamine neurotransmitter systems. These neurotransmitters–serotonin, norepinephrine, and dopamine–are certainly related to mood, but changing the absolute levels of these chemicals in the brain has not been consistently shown to affect mood. Correlation (as I’m sure you’ve heard) isn’t causation.

The most convincing argument against the chemical imbalance hypothesis comes from clinical research into antidepressants. The most commonly prescribed antidepressants are the selective serotonin reuptake inhibitors, also known as the SSRIs. Within hours of taking these drugs, the effective level of serotonin is increased at every serotonin-containing synapse in the brain. But unfortunately, patients who take SSRIs feel no better than patients taking placebo for at least 2 weeks (and realistically, it takes most people 4-6 weeks to really start feeling any effect at all). If depression was caused by a simple chemical imbalance, we’d expect a much more rapid response. And this long time to effect happens with all FDA approved antidepressants, regardless of their action on the serotonin, norepinephrine, or dopamine systems. (I’ve heard some well-meaning doctors tell their patients that the drugs take 4-6 weeks to reach a significant dose in the bloodstream. However, the vastly different half-lives of these drugs and the remarkable consistency in their time to effect argues strongly against this explanation.)

So what causes depression? Depression seems to be related to dysfunctional brain circuitry. A depressed person may very well have problems with one of the monoamine neurotransmitter systems, but it is the system that is impaired, not necessarily the chemical concentration. For recovery to occur, neurons in the brain must find a way to appropriately reconnect with each other.

Thankfully, the brain is remarkably resilient. While as many as 20-40% of the population will experience at least one episode of major depressive disorder, a surprising number recover without any treatment. This isn’t to say that treatment doesn’t work (it does) or that the risk of untreated depression is trivial (it isn’t). It’s just to point out that the brain has mechanisms in place to rebuild dysfunctional synapses. And if you really want to kickstart recover, every known treatment for depression has been shown to increase the levels of neuronal growth factors. The most well known of these factors is brain-derived neurotrophic factor (BDNF), and it is increased by antidepressants, talk therapies, electroconvulsive therapy, and even experimental treatments like ketamine. However, the diverse targets of these therapies may trigger growth in different regions in the brain, which could explain why certain people respond better to one treatment than another. In most cases, the best initial treatment is a combination of antidepressants and cognitive-behavioral therapy.

Why does it matter that we stop talking about depression as a chemical imbalance? For one, an accurate understanding of science matters. When you realize that depression is caused by impaired brain circuitry, it makes sense that depression arises from a combination of genetic, environmental, and psychological factors. On a more human level, calling depression a chemical imbalance trivializes the disorder. A chemical imbalance sounds like something that can be rapidly reversed. It makes us think that there is a magic pill that can quickly fix things. It’s the biological equivalent of the words “why can’t you just snap out of it.” Overcoming depression–especially severe depression–takes time and appropriate treatment. Finally, thinking of depression as a problem with circuitry helps us conceptualize depression as a complex set of disorders with unique causes, manifestations, and treatments. It gives us a reason to approach all sufferers as individuals that cannot be treated with a one-size-fits-all approach.

This model of depression isn’t perfect, and my explanation is, of course, a gross oversimplification. But when we recognize that depression is a circuit problem, it enables us to take some responsibility for rebuilding and maintaining the circuitry that brings us happiness. This is what cognitive-behavioral therapists have been telling us all along. By learning how to combat the thought and behavior patterns that worsen depression, you can help yourself rebuild the faulty circuits within your own head. Conquering depression often takes the help of competent professionals, but recovery without learning to help yourself is almost impossible.

Guidelines for alcohol use

In a recent post, I talked about the importance of behavioral change in preventing early death. This post is the fourth in a series looking at the evidence behind changing health-related behaviors. Previous posts in this series include:

9 surprising facts about weight loss.

Every doctors visit should include smoking cessation counseling.

Medications for smoking cessation


Although heavy drinking is associated with liver failure, brain damage, violence, risky sexual behavior, and motor vehicle accidents, most adults can consume moderate amounts of alcohol without adverse health effects. Some studies even suggest that appropriate levels of alcohol consumption may even be associated with small improvements in certain health outcomes. Of course, there is nothing wrong with being a teetotaler, and weighing the risks and benefits of any alcohol consumption is an important individual decision. If you do choose to drink alcohol, you should consider expert guidelines for appropriate consumption.

In the United States, Dietary Guidelines for Americans recommends no more than 7 standard drinks per week for women, or 14 drinks per week for men. It is also recommends that on any given occasion, women should consume no more than 3 drinks, and that men should consume no more than 4 drinks. There are certain people that should never drink–including pregnant women, children and adolescents, people who cannot limit their drinking, and individuals with certain health conditions (including severe liver disease and illnesses requiring treatment with medications that interact with alcohol). In the United States, one standard drink is considered to be 14 grams of alcohol–the approximate amount found in 12 ounces of beer, 5 ounces of wine, or a 1.5 ounce shot of 80 proof liquor.

Interestingly, drinking recommendations vary significantly around the world. Wikipedia has a really nice summary of recommendations from different countries. In the next few posts, I’ll be looking at some of the evidence behind interventions for people who abuse alcohol.

Why I’m Blogging

I hear a lot of good and bad things about blogging. Some people say it’s a waste of time–and I certainly see how it could be. I started blogging with a modest goal: to enhance my own learning. Writing makes me encapsulate what I learn into meaningful chunks. I can’t promise that what I’ve written will be helpful to anyone else, but it has been extremely useful to me. Now that I have a few posts under my belt, I’ve decided to lay out three main goals from this blog.

1. To learn how to write. My training has focused on technical writing–and I finally feel like I’m becoming comfortable with that style. I love how technical writing emphasizes clear communication of complex concepts. I can’t imagine a better format for writing as a professional. However, my skill at writing to a broader, non-technical audience is much weaker. I struggle to explain what I think with clarity. Hopefully, this blog will help me learn how to write in a crisp, authentic way.

2. To connect with others. There are so many smart people working on difficult problems within the behavioral health community, and I want to learn from them. I believe that there is immense innovation in the pipeline from patients, consumers, doctors, therapists, allied health professionals, academics, and entrepreneurs, and I want to be here as it unfolds. If you’re reading this, I’d love to get to know you better!

3. To advocate for better behavioral health. Behavioral health is misunderstood and often maligned–even within the broader healthcare community. There are many voices for good, and I want to add mine to the conversation. I don’t have many answers, but I am passionate about finding and implementing better solutions.

Medications for smoking cessation

In a recent post, I talked about the importance of behavioral change in preventing early death. This post is the third in a series looking at the evidence behind changing health-related behaviors. Previous posts in this series include:

9 surprising facts about weight loss.

Every doctors visit should include smoking cessation counseling.


As I discussed last week, even brief counseling interventions dramatically increase smoking cessation rates. This effect can be multiplied even further by providing appropriate medications.

There are essentially two main medication strategies for smoking cessation. The first is to replace nicotine from cigarettes with a nicotine-containing gum, lozenge, patch, inhaler or nasal spray. There are some subtle differences between each form, but they all appear to be similarly effective and they increase quit rates by 50-70%.

The second medication strategy for smoking cessation is to use medications such as buproprion, nortriptyline, or varenicline. Bupropion and nortriptyline are considered antidepressant medications, but their effect is probably not related to their antidepressant properties (in fact, other antidepressants have not been shown to be effective for smoking cessation). Both bupropion and nortriptyline produce similar quit rates to nicotine replacement therapies. Varenicline, however, is not an antidepressant. It is thought to work by weakly binding to nicotine receptors, thereby decreasing cravings for nicotine. This is a newer drug, but the early data suggests that varenicline may be the most effective medication yet for smoking cessation.

Every medication has risks and should be discussed with a doctor before starting, but these drugs have been helpful for many people who quit. It would be wonderful if these treatments were offered to more people!