<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[EPD at Digit - Medium]]></title>
        <description><![CDATA[Digit saves, invests, and pays down debt — so you don’t have to think about it. - Medium]]></description>
        <link>https://medium.com/digit-co?source=rss----dbd94a782ce5---4</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>EPD at Digit - Medium</title>
            <link>https://medium.com/digit-co?source=rss----dbd94a782ce5---4</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 04:46:15 GMT</lastBuildDate>
        <atom:link href="https://medium.com/feed/digit-co" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[How we saved a month of engineering time per quarter]]></title>
            <link>https://medium.com/digit-co/how-we-saved-a-month-of-engineering-time-per-quarter-1cc569b0d6c9?source=rss----dbd94a782ce5---4</link>
            <guid isPermaLink="false">https://medium.com/p/1cc569b0d6c9</guid>
            <category><![CDATA[feature-management]]></category>
            <category><![CDATA[fintech]]></category>
            <category><![CDATA[launchdarkly]]></category>
            <category><![CDATA[digit]]></category>
            <category><![CDATA[engineering]]></category>
            <dc:creator><![CDATA[Elaine Lin]]></dc:creator>
            <pubDate>Wed, 13 Jul 2022 18:46:17 GMT</pubDate>
            <atom:updated>2022-07-13T18:46:17.127Z</atom:updated>
            <content:encoded><![CDATA[<p>Replacing our in-house feature management platform with a vendor solution, LaunchDarkly.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*br3SOABrt0qXxtJ8HZknKg.jpeg" /></figure><p>Our previous feature management platform slowed down development. This limited how quickly we could ship new features. Digit engineers spent ~24 days per quarter overcoming limitations with the system:</p><ul><li>~18 days rolling out features to existing users</li><li>~4 days fixing limitations with the platform</li><li>~2 days discussing how to write feature management code</li></ul><p>I experienced these issues myself, and I noticed recurring pain points across engineering. In this post, I’ll talk about how we saved time by replacing our in-house feature management platform with an off-the-shelf solution, LaunchDarkly.</p><h3>In-house feature management</h3><p>Since Digit’s inception, we managed feature flags through an in-house feature management platform. Here’s how it works:</p><ul><li>When a new user signs up for Digit, we create a Digit User model in the database. The User model has a featureFlags field, which is an array of strings.</li><li>You can append/remove feature flags to the User model featureFlags field. To check if a feature is enabled, use the helper function canUseFeature to check if the array contains your feature string.</li></ul><pre>&gt; user.featureFlags<br>[]<br>&gt; user.featureFlags.push(&#39;newSavingsAlgo&#39;)<br>&gt; user.canUseFeature(&#39;newSavingsAlgo&#39;)<br>true</pre><p>This seems reasonable, but there are problems:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rOGr72HzASrDEUshqffhIA.png" /><figcaption>UI for the in-house platform</figcaption></figure><ul><li><strong>Rolling out a feature to existing users requires manual engineering involvement. </strong>The system stores feature flag values in the database. Suppose there are 1 million users, and you want to roll out the feature newSavingsAlgo to 10% of users (100k total). You would need to run a job to update 100k records in the database. Increasing the rollout percentage requires you to run another job. As a feature rolls out, you also need to inform everybody of the rollout status. Each feature ramps over the course of several days (e.g. 1%, 5%, 10%, etc.), so the total overhead cost is ~3 days per feature. In Q4 2021, we had 6 features that affected existing users and required a manual ramp. This adds up to 18 days of engineering time per quarter.</li><li><strong>Gradual feature rollouts are tricky and prone to error. </strong>Feature rollouts lacked guardrails. You could roll out a feature without additional approval or review. This led to incidents. Once, we mistakenly added a feature flag to all users. Fixing the issue required more manual intervention. An engineer ran a job to remove the feature flag and ran another job to re-add the feature flag to the right subset. Each incident is ~3–5 days of overhead — multiple engineers are involved in troubleshooting, rolling back, and postmortem review.</li><li><strong>Writing feature management code is inconsistent and requires additional engineering time. </strong>The in-house system had many ways to manage features: config variables, experiment data models, and several jobs. None of the approaches were robust or easy to use. Figuring out how to write feature management code added ~2 days of overhead per quarter.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Y0fYwpFQGcT2cnpABhQk0g.png" /><figcaption>You are now a Slack bot.</figcaption></figure><h3>Buy vs. build</h3><p>At the end of 2021, we noticed that the in-house feature management system wouldn’t scale with our needs. System issues get worse as you grow. Digit started with a handful of engineers and a savings product. Now, there are more engineers and more products. This increases the number of features we ship each quarter. More features mean more feature management overhead. To be effective, we needed more robust feature management and experimentation tools.</p><p>We could redesign the existing system, but writing code isn’t the only way of solving problems. To fix feature management, we used the buy vs. build framework. Is it better to buy an off-the-shelf solution or build a solution in-house? It can be more efficient to pay somebody instead of writing a complex system yourself. We decided to buy a platform for the following reasons:</p><ul><li><strong>Faster than fixing: </strong>Fixing the existing system would take at least 5–6 weeks for one engineer (1–2 weeks for scoping, 4 weeks for implementation). Integrating an external platform would take about 2 weeks.</li><li><strong>Not our wheelhouse:</strong> At Digit, we focus on improving financial health. Feature management isn’t our area of expertise, and existing vendor solutions fit our use cases.</li><li><strong>Reduced eng involvement.</strong> By buying a platform, we remove most of the eng operational overhead with feature rollouts (about 1 month per quarter). Product managers can ramp features on their own. Engineers don’t need to maintain the feature management platform.</li><li><strong>Similar cost. </strong>The cost of engineering overhead (days spent x cost/day) is similar to the cost of buying a feature management platform.</li></ul><p>After evaluating 13 different vendors, we decided on LaunchDarkly after considering:</p><ul><li>Does the vendor meet our user stories, desired functionality, and security requirements?</li><li>How much effort is needed to integrate the vendor?</li><li>What is the cost of the platform?</li></ul><h3>Out with the old, in with the new</h3><p>Replacing the feature management platform is a big project. It changes how engineers write feature code for all new features. It affects how we manage features for testing, feature rollout, and experiment setup. Here are the core principles we used for integrating LaunchDarkly.</p><p><strong>1. Don’t break functionality.</strong></p><ul><li>Digit has a backend server and mobile apps for iOS and Android. Some users don’t update to the latest app version, so their devices won’t have the most up-to-date code. To address this, we wrote backward-compatible code that doesn’t break older mobile clients.</li><li>There are around 50 legacy feature flags and experiments in the codebase. Cleaning up all of the old code would be risky and expensive. To address this, we took a selective approach to cleanup. We deleted obvious points of confusion and renamed deprecated methods. For example, we prevented creating new experiments on the legacy platform and renamed the method canUseFeature to canUseFeatureDeprecated.</li></ul><p><strong>2. Make the system foolproof.</strong></p><ul><li>LaunchDarkly provides an SDK for evaluating feature flags. If you call the SDK directly, you can accidentally send PII (personally identifiable information) to the platform. To prevent this problem, we added wrappers around the SDK and prevented direct SDK imports. For example, the code looks something like this:</li></ul><pre>await Features.get({<br>  flag: ‘newSavingsAlgo’,<br>  userId: user.id,<br>})</pre><p><strong>3. Introduce the system to the team.</strong></p><ul><li>We trained engineers, product managers, and support ops to use the new system. Engineers learned how to write code in LaunchDarkly. Everyone learned how to use the LaunchDarkly UI. We worked with specific engineers to migrate ongoing projects to the new system.</li><li>Some folks had features that were partially rolled out on the old system. It would be disruptive to migrate these features to LaunchDarkly, so we kept them on the old system.</li></ul><p>It’s now much easier to roll out features at Digit. LaunchDarkly sends automatic Slack messages to notify stakeholders of the rollout process. Product managers can ramp up features without engineering intervention. By improving feature management, we can deliver more features to make financial health effortless for everyone. If you are interested in joining the Digit team, check out our current openings at <a href="https://digit.co/careers">https://digit.co/careers</a>.</p><h3>Acknowledgments</h3><p>Many thanks to Tabitha Blagdon and Michael Brew for reviewing this blog post.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1cc569b0d6c9" width="1" height="1" alt=""><hr><p><a href="https://medium.com/digit-co/how-we-saved-a-month-of-engineering-time-per-quarter-1cc569b0d6c9">How we saved a month of engineering time per quarter</a> was originally published in <a href="https://medium.com/digit-co">EPD at Digit</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Working @ Digit: Elaine Lin, Software Engineer, Onboarding Team]]></title>
            <link>https://medium.com/digit-co/working-digit-elaine-lin-software-engineer-onboarding-team-91fce0f391ef?source=rss----dbd94a782ce5---4</link>
            <guid isPermaLink="false">https://medium.com/p/91fce0f391ef</guid>
            <category><![CDATA[day-in-the-life]]></category>
            <category><![CDATA[software-engineering]]></category>
            <category><![CDATA[fintech]]></category>
            <category><![CDATA[digit]]></category>
            <dc:creator><![CDATA[Elaine Lin]]></dc:creator>
            <pubDate>Wed, 25 May 2022 19:49:16 GMT</pubDate>
            <atom:updated>2022-05-25T19:49:16.006Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_F7aXJ3TepAIfCepKJI3Fg.png" /><figcaption>Spencer Rinkus, Creative Director (left), and Elaine Lin, Software Engineer (right), walk through Digit’s downtown San Francisco offices.</figcaption></figure><p><em>Describe your team’s role at Digit.</em></p><p>I work on the onboarding team at Digit. The onboarding team builds the flows users use to sign up for our products. Onboarding is a critical system because it affects all Digit users. We work hard to make our flow reliable and snazzy so users get a good first impression. Onboarding is also critical for growth and getting new Digit members.</p><p><em>What is one of the most interesting problems your team is trying to solve?</em></p><p>One problem is making sure users get money into Digit. Digit can’t do its job and automate your finances if there’s no money present. We’re working on different ways to fund your account in the onboarding flow. For example, we can have users connect a bank account or encourage them to set up direct deposits.</p><p><em>Walk me through a typical day for you.</em></p><p>Mornings are no-meeting focus time for me. I use this time for focused work i.e. writing code, writing technical documents, reviewing meaty documents. Throughout the day, I check if I can unblock the team. For example, I review code so folks can get their changes out the door, and I pair program to help resolve issues. I have a handful of meetings i.e. 1:1s, document review meetings, company meetings.</p><p><em>What are you currently working on at Digit?</em></p><p>Digit recently got acquired by Oportun, a loan company with affordable loan and credit services. We are working on adding entry points to sign up for Digit in existing Oportun flows. For example, an Oportun user can receive an email with a special Digit promotion. That way, users can benefit from both affordable loans and automated savings.</p><p><em>What are you looking forward to building next?</em></p><p>We plan to replace our rickety in-house feature management system. Incrementally rolling out a feature to existing users requires a multi-day process with eng involvement. I find this very annoying, and I’m excited to replace it. That way, it’ll be quicker and easier to roll out/rollback features. It’ll save a lot of time and frustration.</p><p><em>How is your work at Digit different than other projects you’ve worked on previously?</em></p><p>At Digit, I get to work on a broader range of products and systems. I started out working with credit cards and student loans, and I moved to work on the onboarding flows. I got to learn Go and write backend API wrappers.</p><p>There is a better alignment of priorities. We are efficient with our engineering resources and clear about what we want to ship. If you’re working on a project, you’re likely to see it go out the door.</p><p>The pace is faster. We automatically deploy our server code many times a day. Engineering, product, and design work in tight feedback loops. We can go from idea conception to deployment in just a few weeks.</p><p><em>What is it like working at Digit? [Culture, events, team building, slack vibes, all hands, demos, ERG meetings, etc.]</em></p><p><strong>Culture</strong><br>The work-life balance is good. People respect your PTO time and working hours. My manager encourages taking time off and taking care of my mental health.</p><p>Folks are collaborative. It is easy to get help on issues. When I first joined Digit, I had a lot of questions about Fintech and vocabulary, and I got a lot of helpful answers. When I worked on Go code for the first time, experienced engineers helped me get started.</p><p><strong>Events</strong><br>We had a 3-day hackathon in collaboration with Oportun. Engineers, product managers, and designers worked together on novel ideas. I enjoyed having space to work on dev tooling and improving development speed.</p><p>Teams have social events. While at Digit, I’ve visited a virtual goat farm, putted at mini-golf, and made chocolates. With the pandemic, we’ve gotten good about socializing via Zoom. We’ll play games like Codenames or Pictionary.</p><p><em>What’s your favorite meeting and why?</em></p><p>I run a “Monthly Tech Talks” meeting where engineers can present and share knowledge. It’s a helpful meeting for learning more about new projects, technologies, and ideas. We’ve had people present on Typescript APIs, database configuration, and testing strategies. I like encouraging folks to present their ideas. Public speaking is an important skill, so I’ll help with dry-runs and talk feedback ahead of time.</p><p><em>What’s your favorite slack channel and why?</em></p><p>#dawgz — People post cute dogs.</p><p><em>What was your biggest surprise after joining Digit?</em></p><p>Folks test flows in production. In one case, an engineer sent ~50 internal users each a penny to verify a flow.</p><p><em>What’s been your favorite company or team event?</em></p><p>The Onboarding team had a chocolate-making event. We made chocolate-covered strawberries, dipped Oreos, peanut butter cups, and bear paws. The chocolate was delicious, and they had vegan options!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=91fce0f391ef" width="1" height="1" alt=""><hr><p><a href="https://medium.com/digit-co/working-digit-elaine-lin-software-engineer-onboarding-team-91fce0f391ef">Working @ Digit: Elaine Lin, Software Engineer, Onboarding Team</a> was originally published in <a href="https://medium.com/digit-co">EPD at Digit</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[10 Takeaways from A FinTech Disruption Panel]]></title>
            <link>https://medium.com/digit-co/10-takeaways-from-a-fintech-disruption-panel-b4397cddf8a0?source=rss----dbd94a782ce5---4</link>
            <guid isPermaLink="false">https://medium.com/p/b4397cddf8a0</guid>
            <category><![CDATA[digit]]></category>
            <category><![CDATA[financial-services]]></category>
            <category><![CDATA[fintech]]></category>
            <category><![CDATA[women-in-tech]]></category>
            <category><![CDATA[tech]]></category>
            <dc:creator><![CDATA[Snigdha Kumar]]></dc:creator>
            <pubDate>Thu, 07 Apr 2022 20:10:39 GMT</pubDate>
            <atom:updated>2022-04-11T18:22:40.478Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Digit’s Head of Product Operations attended </em><a href="https://www.linkedin.com/company/association-for-financial-technology-aft/"><em>Association for Financial Technology (AFT)</em></a><em> spring summit and spoke at the FinTech disruption panel. Here are her top takeaways from the panel.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4YxXgf5NKR5U0GWBF96eDQ.png" /><figcaption>Digit’s Head of Product Operations, Snigdha Kumar.</figcaption></figure><p>✅ Even though FinTechs are primarily tech based there is a large focus to build human connection with the users by building personality and moments of delight in the app itself. Customer service is trained to be user centric and user interactions are designed make users feel special.</p><p>✅ Product principles like empathy and simplicity are key to building delightful customer experiences.</p><p>✅ Getting users’ trust is paramount especially in times when everyone is vying for similar users and their paycheck. Having product features that builds trust and provides users control over their finances is key</p><p>✅ Onboarding users is more nuanced that just opening an account, onboarding ends when the users fully understands the product and its value. Engaging with users during their “onboarding” period ensures higher retention</p><p>✅ Growth and Fraud are two sides of the same coin for a lot of FinTechs. If you dial up growth, fraud will rise and that’s the reality of world we live in because of the sophisticated fraudsters and the digital tools at their disposal. To ensure healthy growth with low fraud rates the focus should be on passive data collection (device, phone number, email, etc) and tiered risk levels based on this data. This will allow the FinTechs to focus on seamless customer experience for good users and introduce friction for the fraudulent users.</p><p>✅ KYC/CIP process can no longer be relied upon to truly know your customers because of all the big data breaches that has flooded the dark web with stolen customer identities.</p><p>✅ FinTechs are a wrapper over legacy financial institutions and partners. Right now a FinTech could easily be integrating with 12–15 partners if they want to bring savings, investing and banking product to market. Having multiple partners breaks customer experience, creates too many points of failure and causes customer confusion</p><p>✅ A lot goes into the build vs. buy decision with Financial partners. However, the most important question is whether the partner’s product is core to a FinTech’s business or not. Usually anything that is core to the FinTech’s business model e.g. algorithms, ledger, user interface, is not outsourced to maintain product differentiation irrespective of how good the partner product is.</p><p>✅ Financial partners to FinTech could be a step ahead if they constantly seek feedback on their roadmap from FinTechs. FinTechs are more than happy to do A/B testing of new features or play around with new UIs</p><p>✅ FinTechs are as strong as the financial partners that powers them. Financial partners should design their products by keeping the end customers in mind and proactively add features that are innovative</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b4397cddf8a0" width="1" height="1" alt=""><hr><p><a href="https://medium.com/digit-co/10-takeaways-from-a-fintech-disruption-panel-b4397cddf8a0">10 Takeaways from A FinTech Disruption Panel</a> was originally published in <a href="https://medium.com/digit-co">EPD at Digit</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How We Added Smarts to Funding Your Account]]></title>
            <link>https://medium.com/digit-co/how-we-added-smarts-to-funding-your-account-78c19cd0da6e?source=rss----dbd94a782ce5---4</link>
            <guid isPermaLink="false">https://medium.com/p/78c19cd0da6e</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[neobanks]]></category>
            <category><![CDATA[fintech]]></category>
            <category><![CDATA[digit]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Jaredrfrey]]></dc:creator>
            <pubDate>Thu, 31 Mar 2022 17:50:39 GMT</pubDate>
            <atom:updated>2022-03-31T17:50:38.933Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*WH72Og0Cg2AU30G8" /><figcaption>Jared Frey (author) and Sankeerth Garapati, both members of the AI/ML team at Digit, walk the hallway of Digit HQ in downtown San Francisco.</figcaption></figure><h3>Finding a deposit amount that’s juuust right</h3><p>We’ve spent more than a year creating <a href="https://blog.digit.co/2021/09/16/what-is-direct.html">Digit Direct Powered by MetaBank®</a>, a <a href="https://medium.com/digit-co/how-we-built-a-bank-account-with-a-brain-eedd3eb4d3bf">bank account with a brain</a>. It’s capable of managing your bills, saving for short term goals, and investing.</p><p>Of course, none of this automated money movement is possible if there isn’t any money in a members’ account. Without money, our algorithms couldn’t put anything towards bills or save for a rainy day.</p><p>From user research, we found that many of our members — especially those coming from our original savings product — assumed Digit would take care of moving money from their linked account for them.</p><p>So we listened, then built a feature called “Smart Recurring Deposits.”</p><h3>What are Smart Recurring Deposits?</h3><p>For members with the feature turned on and a linked account — like a checking account from another financial institution — we calculate weekly “smart” amounts to move into Digit. The member is notified about the upcoming deposit and has a day to cancel the transaction before it’s initiated.</p><h3>Finding Goldilocks</h3><p>We want to deposit enough into Digit to let the magic happen, but not more than we need to.</p><p>The first step was to determine the rules for calculating this “Goldilocks” amount. To do this, common first principles of personal finance were our guide: pay your bills, save consistently and have enough money set aside for the day-to-day.</p><p>We look ahead to any upcoming bills and compare that to the balance in your Bills account to calculate a desired withdrawal amount for bills. We then aim to bring your Spending account to a level you’ve set and add a cushion to go towards Savings.</p><p>A separate algorithm actually handles moving money between accounts once it’s in Digit, but this at least ensures that there is enough to work with.</p><p>In our initial testing, some members reported that the amounts were too large. This was an especially common piece of feedback from those coming from our original savings product, which usually only deposits a few bucks per day, rather than larger amounts once a week.</p><p>To mitigate this concern, we modified the algorithm to start small and ramp up. The first deposit would only cover a fixed percentage of the desired withdrawal and was capped at a low dollar amount. That percentage and dollar cap would then increase with each successive deposit until it reached 100% and the overall deposit cap.</p><h3>But first, do no harm</h3><p>Now that we’ve landed on our Goldilocks deposit amount, we still want to make sure this transfer won’t break the bank. Perhaps it’d be great if we could get $215 into Digit, but the connected account only has $200.</p><p>A simple thing to do would be just to draw down to $0, but that could put the user in a tough position if they had upcoming debits out of that account. An even worse idea would be to just try to make the $215 withdrawal and either get declined or cause an overdraft.</p><p>Instead, we want to set a safe upper limit on the Smart Recurring Deposit amount based on the individual’s situation. Our goal is to minimize declined transactions and overdrafts not just for this transfer, but over the next several days.</p><p>We backtested various algorithms using the balance histories for our 600k existing members and calculated the likelihood of the connected account going to $0 or below over the next week. We then settled on an approach that minimizes this likelihood while still depositing the desired amount (or is close and still safe) the majority of the time.</p><p>Finally, we message members the day before the deposit is scheduled, giving them a chance to cancel the transaction with a single tap. As an added precaution, we check the connected account balance one final time right before initiating the transaction to make sure it hasn’t dropped too much since we sent that message. If it has, we don’t initiate the deposit and then let the member know what happened.</p><h3>What’s next?</h3><p>Now that Direct is widely available, more and more members are using this feature. Overall, the response has been positive and we are seeing increased engagement, but the work never ends! We are now seeing heterogeneity in expectations — some members find the amounts too high, some too low, even with the same transfer amount and balances.</p><p>Just like Goldilocks, juuust right isn’t always enough. We’ll keep at it. More to come!</p><p>Direct℠ is a deposit account powered by MetaBank, N.A., Member FDIC.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=78c19cd0da6e" width="1" height="1" alt=""><hr><p><a href="https://medium.com/digit-co/how-we-added-smarts-to-funding-your-account-78c19cd0da6e">How We Added Smarts to Funding Your Account</a> was originally published in <a href="https://medium.com/digit-co">EPD at Digit</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Building investment and retirement goals in Digit]]></title>
            <link>https://medium.com/digit-co/building-investment-and-retirement-goals-in-digit-327ed163b641?source=rss----dbd94a782ce5---4</link>
            <guid isPermaLink="false">https://medium.com/p/327ed163b641</guid>
            <category><![CDATA[investing]]></category>
            <category><![CDATA[retirement]]></category>
            <category><![CDATA[engineering]]></category>
            <category><![CDATA[saving]]></category>
            <category><![CDATA[investment]]></category>
            <dc:creator><![CDATA[Pooja Gada]]></dc:creator>
            <pubDate>Wed, 02 Feb 2022 16:27:00 GMT</pubDate>
            <atom:updated>2022-02-02T16:27:00.627Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Fs6kQGWSJIl8CTv1ib6-ag.jpeg" /><figcaption>A snapshot of the Digit office in downtown San Francisco.</figcaption></figure><p>At Digit we’re on a mission to make financial health effortless for everyone. It’s a huge goal, but like any challenging and impactful problem, it’s achievable by breaking it down and solving it in pieces. One major piece of the financial health puzzle is long-term wealth.</p><p><strong>How investing fits in with Digit</strong><br>It’s been an exciting time for us here lately. We recently launched<a href="https://medium.com/digit-co/how-we-built-a-bank-account-with-a-brain-eedd3eb4d3bf"> Direct — an intelligent bank account</a> that budgets and saves for you. And now Digit can save for investing and retirement goals too.</p><p>Digit gives you time back. It’s one of the things that makes the app so important in our members’ lives. We take the mental math out of saving toward different goals. Digit analyzes your income and spending patterns, then allocates money toward your financial goals. For big goals, like saving toward a down payment on a house or retirement (the ultimate long-term goal!), Digit now offers investing and retirement accounts so members can put their money to work through investments.</p><p>Our goal is to make the investing experience automated and simple. These are the principles we followed while developing these features:</p><ul><li>No account minimums</li><li>Make it easy to understand</li><li>Invest in liquid, diversified, low-cost ETFs</li><li>Investing and retirement features are included as part of a Digit subscription</li><li>Aligned with product vision</li></ul><p>For long-term investments, a portfolio is recommended based on the member’s risk level. Retirement account investments are recommended based on glidepath/proximity to retirement age.</p><ul><li>We recommend a portfolio for you (which you can adjust)</li><li>You can set an optional investing goal target amount</li><li>Digit saves automatically towards your investing goal.</li></ul><p>It literally takes a couple minutes to sign up and be ready to get started.</p><h3><strong>Investing</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/842/1*RnFuxHULL2Am1HShAJukqg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/832/1*4BktHuJPR8Zbad5EtdKjqQ.png" /></figure><h3><strong>Retirement</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/826/1*g0EP4KLplwCnMWIvR9wCGg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/824/1*N8GYEUEMHY-yHcaHc1XdpQ.png" /></figure><p><strong>A closer look at our architecture</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/669/1*ViJxnot6uQG1gPPnbhrFaA.png" /></figure><p>Both investing and retirement goals are powered by our broker-dealer partner <a href="https://drivewealth.com/">DriveWealth</a>. We integrate with them to process all transfers in and out of those accounts and to provide tax information and statements to send to members. For every Digit member that creates an investment goal or a retirement goal, we create a DriveWealth brokerage account.</p><p>Let’s take a peek into the architecture that powers our interactions with DriveWealth.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/751/1*4hBgt6-irOX5SgqkRCPdMQ.png" /></figure><p>Money movement happens either through automatic transfers from our daily savings algorithm or because the member requested a transfer through the app. The application service then creates a transfer and queues a process in our Job Queue system to submit that transfer to DriveWealth.</p><p>Instead of writing those transfers to a file as we do for other ACH transfers, we grab the bank account and routing number from a secure location and send that along in an API request to DriveWealth who then initiates an ACH pull from the user’s linked funding account.</p><p>A difference to note is when a member tries to withdraw most of their funds and how transfers are timed. Normal ACH transfers take roughly 2 business days to settle, but DriveWealth transfers can take upwards of 5 business days. This is because DriveWealth has to convert the member’s assets to cash before they can be withdrawn.</p><p>Another difference is when a member tries to withdraw most or all of their funds. If a member tries to withdraw more than 95% of their funds, we’ll pass along a liquidate flag in the request to DriveWealth. We do this because in the past when we tried to withdraw the value of a member’s portfolio, that transfer would be pending, the market would fluctuate, and then when the transfer was settling it would fail because the user no longer had enough funds to fulfill the original withdrawal amount due to market fluctuations. We picked 95% as the threshold because it’s unlikely for the market to fluctuate by more than 5% before transfers post.</p><p><strong>What else did we build?</strong><br>As you might have guessed, investing requires adherence to strict compliance and controls. We architected a system that would ingest events from DriveWealth (orders, money flow, transactions, etc.) and built monitoring on top of that (ACH transfer delays, incorrect cash transfers, etc.). We monitor whether or not transfers are created successfully and whether or not transfers post successfully, among other important metrics.</p><p><strong>What’s Next?</strong><br>We’re excited to launch investing and retirement features powered by DriveWealth in Digit — it’s a big step towards making financial health effortless. We have an exciting set of additional features on our roadmap that the team is thrilled to build.</p><p>This blog was made possible by contributions from Yubo Diwu— who is an Engineer in the Savings team in product engineering.</p><p>Note: Direct℠ is a deposit account powered by MetaBank®, N.A., Member FDIC.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=327ed163b641" width="1" height="1" alt=""><hr><p><a href="https://medium.com/digit-co/building-investment-and-retirement-goals-in-digit-327ed163b641">Building investment and retirement goals in Digit</a> was originally published in <a href="https://medium.com/digit-co">EPD at Digit</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why I Joined Digit: Jason Palumbo, Marketing Designer]]></title>
            <link>https://medium.com/digit-co/why-i-joined-digit-jason-palumbo-marketing-designer-dd87ca07cf96?source=rss----dbd94a782ce5---4</link>
            <guid isPermaLink="false">https://medium.com/p/dd87ca07cf96</guid>
            <category><![CDATA[design]]></category>
            <category><![CDATA[plants]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[fintech]]></category>
            <category><![CDATA[digit]]></category>
            <dc:creator><![CDATA[Jason Palumbo]]></dc:creator>
            <pubDate>Tue, 23 Nov 2021 19:59:45 GMT</pubDate>
            <atom:updated>2021-11-23T19:59:45.134Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*33Gh9-ZBigkONVhGdR3DmA.jpeg" /><figcaption>Jason Palumbo, Marketing Designer/Musician/Cheese Connoisseur, shares why he made the choice to join Digit in June of 2021.</figcaption></figure><p><strong><em>Tell us about your design journey and how you got to where you are today. Have you always wanted to be a graphic designer?</em></strong></p><p>My design journey started in Edgewater, Maryland when I was taking a high school “photo and digital processing” class. I knew I loved photography already, but there was one project where I had to design a poster with text on top of a photograph that I took advertising our school’s canned food drive. I very quickly became entranced with how typefaces could have an effect on a viewer and how they are used in advertising. I believe it was my teacher, Mrs. Sears, who first told me “You know, you can major in this in college! Look for a graphic design program” …and the rest is history ; )</p><p><strong><em>You’ve been a long-time Digit user, tell us how Digit has played a role in your life today.</em></strong></p><p>I first discovered Digit in 2014 when I was living in Baltimore, MD where I had about 1/4 of the income that I have now. The person who first told me about the product said that it essentially saved money for him “without him even noticing,” which was the perfect thing for a guy liked me who LOVED spending money. I had always struggled with saving because if I knew the money was in my account I would come up with a reason in my head why it was okay for me to spend it. Digit helped me so much just because I could basically set it up then forget about it and it would automatically do the rest based on how and when I was spending. It came in handy especially when I needed new tires for my car (which somehow managed to get nails in them every 5–10 mins). I couldn’t be more excited to not only be a part of telling the world about our new bank account product (Direct powered by MetaBank®) but also be a happy user! I have come a long way since the days where I was scrounging every penny to pay rent, but I am still someone who could probably be more organized with my money and Digit seems like it is the perfect solution for me.</p><p><strong><em>Why did you join Digit?</em></strong></p><p>I believe in the product so deeply!!! As someone who has never felt completely financially healthy being in a creative role and having an income that was on the lower end of the spectrum, I’ve seen how much more confident Digit has made me surrounding my savings. Just knowing that I have a solid rainy day fund (and now, all my bills being paid on time too with Direct) lets me sleep easy at night.</p><p><strong><em>What are you currently working on at Digit? What are you looking forward to designing next?</em></strong></p><p>We are working hard on the design team to get Direct ready to show the world and my role is to explain how it works to people in various types of advertising. It gets me so excited to figure out how to explain it to people because I already love Digit so much. My favorite part so far has been reading some of our member stories and seeing how much of an impact that Digit can have on someone’s life. I am very excited to get some of those stories/images into our ads in the future.</p><p><strong><em>I’ve heard you recently became a flourishing plant parent, what is your favorite plant so far? What advice do you have for aspiring plant parents who may lack a green thumb?</em></strong></p><p>Oh wow, “flourishing,” I am so flattered. I think the best advice that someone gave me was to let the plants do their thing, you aren’t going to kill them overnight unless something crazy happens. I feel that I used to be obsessed with watering them all the time, which lead to problems. Getting a <a href="https://www.amazon.com/XLUX-Soil-Moisture-Sensor-Meter/dp/B014MJ8J2U/ref=sr_1_5?dchild=1&amp;keywords=water+meter&amp;qid=1629241562&amp;sr=8-5">water meter</a> really helped with my plant care because it could tell me how dry the soil was deeper into the pot. This made me less anxious about watering! I highly recommend getting one. Oh and to give you an answer to the part about my favorite, how could I ever choose one of my babies as a “favorite” when they are all so dang gorgeous!?….. (The huge monstera is by far my favorite. I love her so much and I’ve been growing her into a gorgeous adult for 2 years. Don’t tell the other kids.)</p><p><strong><em>If you had to describe yourself as a gif, which one would you use?</em></strong></p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fgiphy.com%2Fembed%2Fhs7AIkexF2PX4kuWxQ%2Fvideo&amp;display_name=GIPHY&amp;url=https%3A%2F%2Fgiphy.com%2Fclips%2Frupaulsdragrace-drag-race-dragrace-anne-hathaway-hs7AIkexF2PX4kuWxQ&amp;image=https%3A%2F%2Fmedia3.giphy.com%2Fmedia%2Fhs7AIkexF2PX4kuWxQ%2Fgiphy.gif&amp;key=d04bfffea46d4aeda930ec88cc64b87c&amp;type=text%2Fhtml&amp;schema=giphy" width="480" height="270" frameborder="0" scrolling="no"><a href="https://medium.com/media/87908656ad66f0adda34849b04df277b/href">https://medium.com/media/87908656ad66f0adda34849b04df277b/href</a></iframe><p><strong><em>What’s one piece of design advice you would give to your younger self or even new designers out there?</em></strong></p><p>Learn to take negative AND positive feedback, because you’re going to get a lot of both during your career. Being able to take compliments is just as important as being able to receive criticism about your work. In terms of the latter, an important thing I had to learn is to make sure I was only making intentional design decisions. If you disagree with someone’s critique of your work you won’t be able to make a case for the choice you made unless there was some intentionality behind your design. Everyone is going to have different opinions and reactions to a piece of work, but having a strong set of reasons that you can explain clearly to anyone makes more people respect your choices.</p><p>Note: Direct℠ is a deposit account powered by MetaBank®, N.A., Member FDIC.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=dd87ca07cf96" width="1" height="1" alt=""><hr><p><a href="https://medium.com/digit-co/why-i-joined-digit-jason-palumbo-marketing-designer-dd87ca07cf96">Why I Joined Digit: Jason Palumbo, Marketing Designer</a> was originally published in <a href="https://medium.com/digit-co">EPD at Digit</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How we scaled user research for the Digit Beta]]></title>
            <link>https://medium.com/digit-co/how-scaled-user-research-for-the-digit-beta-7a712df2457b?source=rss----dbd94a782ce5---4</link>
            <guid isPermaLink="false">https://medium.com/p/7a712df2457b</guid>
            <category><![CDATA[democratization]]></category>
            <category><![CDATA[ux-research-process]]></category>
            <category><![CDATA[ux-research]]></category>
            <category><![CDATA[user-research]]></category>
            <category><![CDATA[fintech]]></category>
            <dc:creator><![CDATA[Clara Kuo]]></dc:creator>
            <pubDate>Wed, 03 Nov 2021 20:10:30 GMT</pubDate>
            <atom:updated>2022-05-19T23:46:36.209Z</atom:updated>
            <content:encoded><![CDATA[<h3>How we scaled user research for the early Digit Beta</h3><h4>Also: how we reached 100 member exposure hours!</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fUMMJj-w6fLVI8dRw9upAQ.jpeg" /><figcaption>Clara Kuo, Lead User Researcher, speaks to Gunjan Kushwaha, Engineering Manager.</figcaption></figure><p><strong>Why we care about user feedback at Digit</strong><br>For most user researchers, democratizing research can be a challenge. How do we navigate the resistance? Also, how do we help PMs leverage member feedback as part of the product roadmap?</p><p>I’m proud to say that our team at Digit invested as many as 100 hours in the first quarter of the Digit Beta to listen to our users, and it was a team effort.</p><p>We have a culture of internal feedback that speaks directly to one of our core values: <em>Build Trust.</em> This same core value also applies to our Digit members. We aim to share knowledge about our members as a team, regardless of our job roles or titles.</p><ol><li>Key Learnings</li><li>Shared Insights in Real Time</li><li>Research as a Team</li><li>Templates &amp; Reports</li></ol><h4><strong>Key Learnings from creating Democratized Research</strong></h4><ul><li><strong>UX objectives come before UX methods:</strong> “Which method and when to use?” is arguably the most classic UX research FAQ. But the practice of leading with a method first is outdated. Define the key research objective first and how it should inform a stakeholder’s business decisions. Your team will find the research more relevant to them. If we focus on developing UX strategy and defining the right outcomes and impact for the org, our discipline will grow with the business.</li><li><strong>Define research roles: </strong>In a scrappy startup like Digit, sometimes stakeholders will lead studies where the researcher acts as a “co-navigator”. Other times, the researcher acts as the “driver.” For each study, it’s important that we clearly define the roles. When stakeholders lead their own user research studies, they develop more appreciation of user problems and problem solving</li><li><strong>Rome was not built in a day.</strong> When I started on this journey with my product partners, it had its natural ups and downs. There were a lot of questions how research should run, and they were really interested in how they could get involved. This is a form of feedback and it showed that our team cared about the outcomes. If you’re a researcher trying to enable others to run research, help your team understand that how research efforts can help yield incredible results!</li></ul><p><strong>Shared insights in real time </strong><br>At Digit, we have a specific Slack channel for beta triage with product, engineering, and support. When research is in flight, we share research updates and audio clips in our Slack channel for wider exposure. To help the team see the flow of user feedback in a more complete form, we tracked feature feedback in real-time from support &amp; research sources and created a kanban board of feedback.</p><p>For the initial beta period, this has helped increase the overall awareness of our research internally. Stakeholders are able to decide how and when to implement potential solutions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*233BXaVyxlWkXb2tcI_QFg.png" /><figcaption>How we track user feedback in a UX backlog</figcaption></figure><p><strong>Research as a Team</strong><br>When I joined Digit last November, our product &amp; design teams were already speaking directly to users. Early on, they shared that hearing user ideas, pain points and needs is valuable. But, every study could run differently, based on who was running it.</p><p>Since January, PMs and designers work with Research and rely on us to make running a research study seamless. PMs and designers can request <em>Digit Chats</em> and conduct user interviews with light research guidance. We tracked user participation, troubleshooted no-shows and managed post-interview incentives to make it easier for the team.</p><p>We also try to recognize our Digit teammates for their research efforts. We call these <em>Digits doing Research</em> shoutouts and post to Slack in weekly updates.</p><p><strong>Templates &amp; Reports</strong><br>When Digit stakeholders conduct research themselves, they hear about user pain points directly. When they co-write research findings, PMs or designers may consider them in their product requirements.</p><figure><img alt="Research Template Outline" src="https://cdn-images-1.medium.com/max/329/1*w8cziQwhvBxHl9jA_vyNLA.png" /><figcaption>Template: Run a research study the same way every time</figcaption></figure><p>Building our “empathy muscle” helps our team to refine their decisions, guided by user feedback. It’s rewarding to co-create research findings with them and help clarify their recommendations.</p><p>Today, we use Coda.io to host a single, consistent template for creating research insights for the Digit Beta. It also helps us so that we can find any research related to the Digit Beta in a single location.</p><p>Stakeholders like to see a Slack post of the TL;DR or short summary of the research findings with a link to the Coda. They also like to listen to clips as evidence. For most studies I’ve led at Digit, I’ve been able to create almost zero Powerpoint decks for sharing insights (a personal goal)!</p><p><strong>In closing… </strong>That’s the magic of Digit. Because our team is focused on impact, it helps to be part of a larger journey and to know we are helping users together.</p><p><em>Note: Direct℠ is a deposit account powered by MetaBank®, N.A., Member FDIC.</em></p><p><strong>UX Acknowledgments:</strong> Shipra Kayan, Emily Bergeron and Mariné Palmuyatan for a lively debate early this year about Voice of Customer programs; and Brigette Metzler and the ReOps community</p><p><strong>Digit Acknowledgments:</strong> Thanks to the PM, Marketing, Design and Analytics teams for your endless feedback.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7a712df2457b" width="1" height="1" alt=""><hr><p><a href="https://medium.com/digit-co/how-scaled-user-research-for-the-digit-beta-7a712df2457b">How we scaled user research for the Digit Beta</a> was originally published in <a href="https://medium.com/digit-co">EPD at Digit</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Roadmapping Towards Effortless Financial Health]]></title>
            <link>https://medium.com/digit-co/roadmapping-towards-effortless-financial-health-824c462a6f51?source=rss----dbd94a782ce5---4</link>
            <guid isPermaLink="false">https://medium.com/p/824c462a6f51</guid>
            <category><![CDATA[roadmaps]]></category>
            <category><![CDATA[product]]></category>
            <category><![CDATA[fintech]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[digit]]></category>
            <dc:creator><![CDATA[Samar Shah]]></dc:creator>
            <pubDate>Tue, 19 Oct 2021 18:40:35 GMT</pubDate>
            <atom:updated>2021-10-19T18:40:35.252Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*U29gTLuoCGAGwLvUZ1u3MQ.jpeg" /></figure><p>The below questions were shared with Built In for a story published in July of 2021 focused on <a href="https://www.builtinsf.com/2021/07/29/weathering-change-product-roadmap">Weathering Change on a Product Roadmap</a>. Our Head of Product, Samar Shah, shared how he and his teams stay in sync:</p><p><strong>Interview Questions — Answer here</strong><br><strong><em>1. When developing a product roadmap, what steps do you take to ensure there’s alignment across teams from the get-go?</em></strong></p><p>The first step is making sure everyone understands how our mission translates to our product strategy. At Digit, our mission is “To Make Financial Health Effortless for Everyone”. We intentionally opted for a business model that transparently maps to this mission: our customers pay a monthly subscription for Digit’s automated financial health platform.</p><p>This then drives our product strategy: to deeply understand the most important needs in our customers’ financial lives, build a roadmap of features that intelligently solve these needs with Digit’s effortless automation, and ultimately increase the impact our members receive from our subscription.</p><p>From there, we kick off each planning cycle by aligning on companywide product goals: a portfolio of metric-based and milestone goals to deliver product and business impact. We share out draft versions of goals early on to allow time to strengthen them based on cross-functional feedback.</p><p>Once everyone is aligned on our companywide strategy and goals, we rely upon each product team’s expertise to chart the roadmap ahead. Digit product teams are made up of dedicated product managers, engineers, designers, data scientists, and business stakeholders. Each team owns a different mandate and domain within our product experience and they work together to ensure that our collective roadmap will ladder up to reach our company goals.</p><p><strong><em>2. How do you maintain that alignment throughout the development cycle? Cite a specific example of this process in motion.</em></strong></p><p>I’ve found it most important to set up all teams to transparently communicate their progress and obstacles. If this information is readily available and well-consumed, teams will naturally lean in to support each other and adapt to the latest learnings.</p><p>We use Coda to maintain a collaborative hub that houses our “master product roadmap” and the status of each initiative. We operate our roadmap as a living, breathing stack rank, so mid-cycle changes to priority are immediately reflected. Dynamically linked team-level views allow each of our product managers and team leads to drive their own team-level planning, while automatically keeping the master roadmap in sync on the latest status.</p><p>Even then, it can be overwhelming to stay up to speed with everything in flight. To help with this information overload, we run a weekly heartbeat meeting that is broadly attended and allows each team to share their most important priorities and how they are trending towards their goals.</p><p><strong><em>3. As project needs change, how do you re-prioritize the product roadmap and keep teams aligned?</em></strong></p><p>Roadmap reprioritization usually occurs through a series of decisions based on new information: “We just learned X, so we should take Y action”. At Digit, our favorite way of getting from X to Y is with a written memo.</p><p>A well-written memo can lay out the new learning, how it changes our prior strategic assumptions (e.g., “early data suggests a gap in how our members are using a new feature” or “we underestimated the technical complexity of this project”), and how it impacts our ability to reach our goals. In most cases, we can still reach our goal if we shuffle prioritization or adjust team resourcing. Less frequently, the latest learnings may require us to revise our goals.</p><p>We ask our project leads to outline the different decision options and make a recommendation. This allows stakeholders to read and deeply understand others’ thought processes. We believe that async processing leads to stronger thinking and levels the playing field for everyone to express their opinions, so we wait to gather written feedback before holding live discussions. Similarly, we don’t force decisions in meetings and let project leads break from a meeting when they have enough input to ‘sleep on it’ and make a decision.</p><p><strong><em>Interested in joining our team and working on our Financial Health roadmap? Reach out to </em></strong><a href="http://recruiting@digit.co/"><strong><em>recruiting@digit.co</em></strong></a><strong><em>.</em></strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=824c462a6f51" width="1" height="1" alt=""><hr><p><a href="https://medium.com/digit-co/roadmapping-towards-effortless-financial-health-824c462a6f51">Roadmapping Towards Effortless Financial Health</a> was originally published in <a href="https://medium.com/digit-co">EPD at Digit</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Digit’s Guide to Making Decisions]]></title>
            <link>https://medium.com/digit-co/digits-guide-to-making-decisions-9c3549ddb53c?source=rss----dbd94a782ce5---4</link>
            <guid isPermaLink="false">https://medium.com/p/9c3549ddb53c</guid>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[coda]]></category>
            <category><![CDATA[digit]]></category>
            <category><![CDATA[saving-money]]></category>
            <category><![CDATA[decision-making]]></category>
            <dc:creator><![CDATA[Snigdha Kumar]]></dc:creator>
            <pubDate>Fri, 17 Sep 2021 16:44:45 GMT</pubDate>
            <atom:updated>2021-09-18T13:40:25.229Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Is3D3JrCEfERU448tRdDkg.jpeg" /></figure><p>Use our <a href="https://coda.io/@snigdha-kumar/digits-guide-to-making-decisions">Decision Template</a> to make great decisions!</p><h3>Introduction to Digit</h3><p>Most Americans don’t have enough money put away for a rainy day. At <a href="http://digit.co/">Digit</a>, we’re working to change that. We want to eliminate the stress and anxiety people feel about their finances so they can focus on what’s most important in their lives. We first mastered saving for people’s near-term goals, helping our members automatically save over $5 billion. Now, we are harnessing our groundbreaking technology to give our members an automated smart financial assistant to manage all of their personal finances.</p><h3>Making Decisions</h3><p>We believe in making good, durable decisions at Digit. Durable decisions should remain unchanged unless key inputs or assumptions change. We may get some decisions wrong; that is okay as long as we understand the context behind our decision, what we got wrong, and how we could do better in the future.</p><p>We also believe in transparency and rigor. We achieve these goals, with the aim of making durable decisions, through written documents, asynchronous debate, and public posts about decisions and their follow ups.</p><p>We are believers in the <a href="https://coda.io/d/Digits-Guide-to-Making-Decisions_d-a0nyoLhta/DACI_suJln">DACI framework</a>. In particular, this document is aimed at Drivers of a decision.</p><p>The short process of making a decision at Digit:</p><ol><li>The driver writes a decision document and shares with stakeholders.</li><li>We review the decision asynchronously through slack or in “office hours”: a meeting with key stakeholders and our executive team.</li><li>We either reach a decision or ask for follow-up information. In the latter case, we repeat.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*JLb4iQl8TJ6go6m6" /></figure><p>Not all decisions require this degree of rigor. We want to empower individual teams to make their own calls, especially for reversible decisions. We <em>require</em> this degree of rigor for Irreversible decisions and expect to spend time on them as a company.</p><p>This document is aimed at detailing out all the steps that enable decision making at Digit.</p><h3>1. Write a ‘good’ decision document</h3><p>Digit has a strong documentation culture. We use documents as our primary communication tool for recording facts, laying out options, driving decisions. Writing a good document requires thinking about 3 main components:</p><p>A. Document structure</p><p>B. Document substance</p><p>C. Document formatting</p><h3>A. Document structure</h3><p><em>Jump to the </em><a href="https://coda.io/d/Digits-Guide-to-Making-Decisions_d-a0nyoLhta/Decision-Template_suWCj"><em>Decision Template</em></a></p><h3>Start with a clear and succinct recommendation.</h3><p>This lays out where you stand on a particular decision and does not leave the reader hanging.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*OgutN8QwNg4vUv5e" /></figure><h3>Have a “goal/ non-goal section” upfront</h3><p>A goals section makes it clear what the meeting or review is about. This helps us maintain focus and keep things on track.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ekMuD8Ym2FZEeVLp" /></figure><h3>Make the DACI explicit</h3><p>Upfront DACI aids in decision making and helps participants know their roles. Read more in <a href="https://coda.io/d/Digits-Guide-to-Making-Decisions_d-a0nyoLhta/DACI_suJln">DACI</a> .</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/670/0*ghKja9YcNpBxMs_K" /></figure><h3>Explain the What, Why, and How</h3><p><em>What </em>is a factual understanding of the world. We lay the groundwork by bringing everyone up to speed on what we know.</p><p><em>Why</em> are the reasons that explain the what.</p><p><em>How</em> is our methods or alternatives to fix the problem.</p><p>Mindmaps are a great tool for a visual representation of complex problems.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*UrnZCjTsOF4fcxaS" /></figure><p>Further reading: [<a href="https://rushabhdoshi.com/posts/2021-04-19-complex-problem-solving/">blog post</a>, <a href="https://www.amazon.com/Strategic-Thinking-Complex-Problem-Solving-ebook/dp/B01HS3YCM8">book</a>]</p><h3>Have a feedback section in the end for questions/ discussion</h3><p>Long debates in comment threads are difficult. Our way of having these discussions is to create a section at the end of the document where people can put in their questions, get them answered, or comment on others’ questions and have a right, asynchronous discussion. This also helps record all the viewpoints in one place to be reflected later incase of a change in decision.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*STsDLCGNiNWkg-1d" /></figure><p><strong><em>You can copy the template to make your own decision documents </em></strong><a href="https://coda.io/d/Digits-Guide-to-Making-Decisions_d-a0nyoLhta/Decision-Template_suWCj"><strong><em>here</em></strong></a></p><h3>B. Document substance</h3><h3>Present all the options considered</h3><p>Bring your audience along by showing all the options considered. Tables are a good way of representing two dimensional data. Narrative form is an acceptable form as well.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*_HTM1OX3_WvuJxPk" /></figure><h3>Add supporting data and analysis to back-up the recommendation</h3><p>Any recommendation should be followed by data and analysis that went into making that recommendation. Analysis can be quantitative or qualitative (e.g. reference checks, industry benchmarking, user feedback, and user studies.)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*xjDah3IGje_f2nUs" /></figure><h3>Key stakeholders should have vetted the recommendation</h3><p>For example, if your decisions needs engineering to implement something, relevant engineers should not be surprised in an office hours review or find out about your recommendation is made. Additionally, when decisions require work from other teams, do the homework of estimating the time or complexity of the demand. As a simple rule of thumb make sure that all the “contributors” from your <a href="https://coda.io/d/Digits-Guide-to-Making-Decisions_d-a0nyoLhta/DACI_suJln">DACI</a> have vetted the recommendation.</p><h3>C. Document formatting</h3><h3>Use heading hierarchy and sections and separate out a large document</h3><p>Using proper H1, H2, H3 formatting creates automatic indexes in Coda. It makes our documents consistent and easier to read.</p><h3>Use colors judiciously</h3><p>Colors can help highlight complicated tables or parts of dense documents. However, remember that the focus is on making a <em>good</em> decision, not on selling <em>your </em>option.</p><p>A good use of colors is to highlight cells in a table that are positive, neutral, and negative.</p><p>A poor use of colors is highlighting entire paragraphs so that they stand out in contrast to others. This significantly reduces readability.</p><h3>Use tables, miro charts, images</h3><p>The goal is to help people understand what you know and your reasoning. Visual aids can make a significant positive difference. Use all the tools at your disposal to easily grasp the document.</p><h3>2. Sharing and Review</h3><p>Once you have written out a decision document, the next step is sharing it with the decision makers and contributors, soliciting feedback, and making a decision.</p><p>All decisions require async sharing. Some decisions may need a live meeting. We leave it to the Driver’s judgement to request a live meeting. A lot of back and forth commenting and feedback in your doc is a good sign that you should probably do a live meeting to get everyone on the same page.</p><p>We prefer doing things async. The general rule of thumb if the decision is simple, doesn’t require a lot of discussion and there is already stakeholder buy-in for the recommendation then use the async route for document review.</p><p>While there are multiple styles of running office hours, some common do’s and don’ts of running a successful outline are presented below</p><h3>Send your decision document at least 1 day before your meeting</h3><p>This gives people ample time to read and internalize the doc, and activate their slow brain. If you’re not seeing sufficient feedback, nudge key participants to add feedback and questions ahead of the meeting. We use a simple text feedback section at the bottom of the doc.</p><h3>Prepare</h3><p>Answer as many simple questions as you can async. Read through the more complex questions and focus on the contentious or difficult feedback and questions in the live meeting.</p><h3>Keep the meeting focused</h3><p>In any live meeting, people will bring up ancillary points and the discussion may meander toward things that are tangential or are non-goals. Your job is to keep the discussion focused on the goals of the meeting.</p><h3>Push the ‘approver’ to come to a decision or ask for clear follow-ups/ action items</h3><p>Ensure you don’t leave the meeting without a clear decision or a clear set of follow up tasks.</p><h3>Don’t be married to your recommendation</h3><p>Set your ego aside to reach the best decision for Digit, even if the decision is different from what you initially suggested.</p><h3>3. Record the decision</h3><p>Outcome of an office hour or async review is usually either a follow-up request for more information or a final decision.</p><p>Once you reach the stage of final decision, update your document with the final decision. Share a summary of the final decision in the #officehours slack channel for broader visibility.</p><h3>Author</h3><p><a href="https://www.linkedin.com/in/snigdha-kumar-00620742/">Snigdha Kumar</a> is the Head of Product Operations at Digit. She has spent most of her professional life helping increase access to financial services through technology. She is interested in all things FinTech innovation and strategy.</p><h3>Acknowledgements</h3><p>This work would not have come together without the contribution and guidance of <a href="https://www.linkedin.com/in/radoshi/">Rushabh Doshi</a> (Chief Product Officer @ Digit), <a href="https://www.linkedin.com/in/ryannier/">Ryan Nier</a> (General Counsel @Digit) and <a href="https://www.linkedin.com/in/davidkossnick/">David Kossnick</a> (Product leader at Coda). Thank you for your all your edits, suggestions and review.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9c3549ddb53c" width="1" height="1" alt=""><hr><p><a href="https://medium.com/digit-co/digits-guide-to-making-decisions-9c3549ddb53c">Digit’s Guide to Making Decisions</a> was originally published in <a href="https://medium.com/digit-co">EPD at Digit</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[To incentivize or not to incentivize — is it worth the cost?]]></title>
            <link>https://medium.com/digit-co/to-incentivize-or-not-to-incentivize-is-it-worth-the-cost-ded5cfc4946c?source=rss----dbd94a782ce5---4</link>
            <guid isPermaLink="false">https://medium.com/p/ded5cfc4946c</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[data-science-at-digit]]></category>
            <category><![CDATA[fintech]]></category>
            <dc:creator><![CDATA[Tessa Johnson]]></dc:creator>
            <pubDate>Tue, 27 Jul 2021 22:04:02 GMT</pubDate>
            <atom:updated>2021-07-27T22:04:02.590Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>To incentivize or not to incentivize — is it worth the cost?</strong></h3><p><strong>Decision frameworks to analyze the effectiveness of incentives</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KvuON7TPPfhSE8-Hrj0fMg.jpeg" /></figure><p>At Digit we love to ship new features that help our members achieve financial health. While we always believe that these features are beneficial, it’s a natural part of the member lifecycle for a change to cause a subset of members to leave the platform (churn). One way we mitigate churn is with incentives, such as offering members a free month on the platform or another kind of promotion. These incentives can be helpful tools that can keep members on the platform long enough to recognize that a product improvement is valuable.</p><p>Incentives come with a cost. We want to know overall if the cost of the incentive will yield a positive result for Digit. After all, we want to remain financially healthy as a company, too. We can explore this from a data science perspective in two different ways:</p><ul><li>Long term revenue goals: Is the revenue gained from churn reduction greater than the cost of the incentive?</li><li>Short term customer acquisition goals: Is the cost of preventing a member from churning less than the cost to acquire a new member?</li></ul><p>In order to answer these questions, we perform an A/B test to measure the churn levels that will feed into our econometric models.</p><p>This post is about setting up a decision framework <em>before</em> initiating an experiment to measure churn. We can use it to determine the acceptable combinations of churn that justify the cost of the incentive. The decision framework also helps us set the sample sizes for the experiment.</p><p>We’ll share some formulas and code to help along the way.</p><h3>A/B test setup</h3><p>The initial A/B experiment has two purposes. The first is to discover the level of churn that will occur with the new feature rollout. The second is to measure the percent reduction in that churn when an incentive is offered with the feature rollout. If we are testing multiple incentives, they can each have their own variant in the test:</p><ul><li>Variant A: new feature rollout, no incentive</li><li>Variant B: new feature rollout, with incentive X</li><li>Variant C: new feature rollout, with incentive Y</li><li>etc.</li></ul><p>The percent reduction in the churn (<em>R</em>) is a negative relative lift between the churn from variant A (<em>Cᴀ</em>) and churn from variant B (<em>Cʙ</em>).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/124/1*1mcAMigB8cfLUHMjEgPN2Q.gif" /></figure><p>For example: If variant A has a churn rate of 35%, and the incentive offered to variant B reduces that churn rate to 30%, <em>R</em>=14.3%.</p><p>The sample sizes for the experiment will be determined <em>after</em> analyzing our chosen decision framework.</p><h3>Decision frameworks</h3><p>Before we begin the experiment, we want to know which combinations of churn from variants A and B (and the calculated reduction in churn, <em>R</em>, between them) will result in a positive return on our investment of the incentive. We do this by listing out all possible scenarios and what our corresponding payout would be.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NQVE3rs5ShHY8HedoVb97Q.png" /></figure><p>The next two sections explore churn combinations that would justify an incentive for both long-term and short-term business goals.</p><h4><strong>Long-term goal: do we ever make a positive return on the investment of the incentive?</strong></h4><p>Digit’s business is subscription based, and we can project out the revenue expected from a cohort over time, including churn that would naturally occur. To understand the effect of different combinations of churn, we can look at the projection of our revenue as a function of time in the case where a new feature rollout caused churn, and in the case in which an incentive prevented some members from churning, but came with an up-front cost.</p><p>In the case of no incentive, revenue after M months is calculated as:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/335/1*4oUsTckJzREQRLWFBNboDg.gif" /></figure><p>where <em>N</em> is the number of members in the cohort before the feature rollout, <em>S</em> is the monthly cost of the subscription, <em>Cᴀ</em> is the churn from variant A, and <em>Cɴ</em> is the natural churn level of the cohort each month without the introduction of any feature changes.</p><p>For the case where there is an incentive, we may have saved some members (from our definition of <em>R</em> in the section above, the percent churn becomes <em>Cʙ=Cᴀ*(1-R)</em>), but we have to subtract off the cost of the incentive, <em>I</em>,</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/449/1*dM6bThULE--n450sGAbxsw.gif" /></figure><p>The revenue gained from the incentive after <em>M</em> months would be</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/140/1*AN01ZxIWKSSqj0aVPrB1sA.gif" /></figure><p>Using python we can code up a visualization of the revenue gained from the incentive after a certain number of months for different combinations of <em>Cᴀ</em> and R. The toy example (code and resulting figure) below looks at the revenue saved after 12 months assuming that there are 50,000 initial customers in the cohort, the subscription cost is $3 per month, there is a natural churn level in the cohort of 5% each month, and the incentive costs the company a total of $100,000.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/23eabea0c5bafa37b0edb347ab3c5191/href">https://medium.com/media/23eabea0c5bafa37b0edb347ab3c5191/href</a></iframe><figure><img alt="" src="https://cdn-images-1.medium.com/max/895/1*ne5gByJKPo9_k4FG-pg0kQ.png" /></figure><p>In this toy example case, if group A from the A/B test shows a churn of 5%, it doesn’t matter what <em>R</em> is — 12 months after offering the incentive, the company will have lost money from the cost of the incentive when few members were going to churn anyway. If churn is very high but <em>R</em> is low, like the case of <em>Cᴀ</em> = 40% and <em>R</em>= 10%, the incentive didn’t prevent very many members from churning and revenue is lost. The long-term outlook for the incentive is only profitable if both <em>Cᴀ</em> and <em>R</em> are high, like in the case of <em>Cᴀ</em> = 30% and <em>R</em>= 40%.</p><p>Here, we would choose minimum sample sizes for an A/B test that would provide sensitivity to combinations of <em>Cᴀ</em> and <em>R</em> that provide a positive long-term revenue projection.</p><p>**Note — there are cases that could complicate the formula. For instance, if an incentive is offered only to non-churned members, the overall cost of the incentive will be dependent on the churn level of variant B. We won’t discuss those cases here.</p><h4><strong>Short-term goal: do we see comparable cost per acquisition to a new member?</strong></h4><p>If a company is actively marketing to gain new members, it is likely that there is an accepted cost per acquisition, or CPA (how much money on average a company is willing to spend to acquire a new member). If churn is expected due to a change in the product, the cost of an incentive which prevents some members from churning can be used to calculate a unit cost per saved member, analogous to the CPA. We can find out what the cost per saved member is as a function of <em>Cᴀ</em> and <em>R</em> from our A/B test and compare to our accepted CPA.</p><p>If the cohort that will see the new feature has <em>N</em> members, and we have measured churn levels <em>Cᴀ</em> and <em>Cʙ</em> from our A/B test described above, the number of members saved from churning, <em>Ns</em>, can be calculated as follows:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/225/1*X0M4RIe22vXm6PWDB5OnOg.gif" /></figure><p>using our previous formula for the reduction in churn, <em>R</em>. This would give us a cost per acquisition of:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/122/1*bl6Ei3qkbYK4YX50luXBLg.gif" /></figure><p>where I is the cost of the incentive.</p><p>Again, before even beginning an A/B test, we can develop a framework for understanding the result, which will assist in choosing the minimum sample sizes needed to see an effect that would justify the cost of the incentive.</p><p>The python code below creates a figure that shows cost per saved member (CPA, if being treated as a cost per acquisition) as a function of<em> Cᴀ</em>, given different levels of <em>R</em>. A horizontal line can be drawn across the accepted cost per acquisition to find the combinations of <em>Cᴀ</em> and <em>R</em> that would result in the incentive being justified.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/566064653f7bb479df6a2251ae826e0d/href">https://medium.com/media/566064653f7bb479df6a2251ae826e0d/href</a></iframe><figure><img alt="" src="https://cdn-images-1.medium.com/max/853/1*RNeK0cW8i4Qq7AmjGOUT6g.png" /></figure><p>** Note — the value of an existing user (measured by future lifetime value, perhaps) who is saved from churning may be greater than a new user who has never tried the product! The cost per saved user that a company is willing to accept in offering an incentive may be greater than the cost per newly acquired user based on the projected value of the two different cohorts.</p><h3><strong>Making decisions and the order of operations</strong></h3><p>As data scientists, it is our job to support decision making at the company through data, modeling, and experimentation. Through these decision making frameworks, we can help the company decide what goals we are trying to achieve by offering incentives and then set up the least invasive experiment possible to measure whether or not we will achieve those goals.</p><p>The smoothness of decision making around offering incentives is dependent on our order of operations. We want to work in this order:</p><ul><li>determine a list of potential incentives and estimate their cost,</li><li>determine whether the goal is long-term revenue gain or short-term member “acquisition”,</li><li>calculate the combinations of churn from an A/B test that would achieve the company goals,</li><li>determine the minimum experiment sample size that would provide sensitivity to combinations of churn that would justify offering the incentive, and</li><li>set up and run the experiment.</li></ul><p>If we work in this order, after the experiment finishes running and the results are in, we already know what to do. If we detect a significant reduction in churn from the variant that is offered the incentive, we can compare to our long-term or short-term goal charts to determine if it is worth the cost.</p><h3>Acknowledgments 💙</h3><p>Thank you to Ravi Soni (Data Scientist at Digit) for doing this work with me!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ded5cfc4946c" width="1" height="1" alt=""><hr><p><a href="https://medium.com/digit-co/to-incentivize-or-not-to-incentivize-is-it-worth-the-cost-ded5cfc4946c">To incentivize or not to incentivize — is it worth the cost?</a> was originally published in <a href="https://medium.com/digit-co">EPD at Digit</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>