PETRA NOVANDI BARUS Lean Startup Engineering 

Petra Novandi Chief Technology Officer Kuncie.com petrabarus petrabarus KodingBarengPetra 

My Startup Journey Co-found UrbanIndo.com 2011 Nov 1st Fundraising 2012 2nd Fundraising 2013 Largest Real-estate Portal Indonesia 2014 Acquired by 99.co 2018 Joined Amazon Web Services 2019 Joined Kuncie 2021 1million users 2021 

Common Startup Pitfalls 

1 Overestimating Market Size 2 Poor Market Research 3Design Problems 4Excessive Development Costs 5Poor Customer Acceptance 6Competitive Reaction Why Product Fails 

a stellar product is not built using one big leap, but through small incremental steps.... 

1 Quick fast experiments leads to answers 2 Determine Problem/Solution Fit with MVP 3Iterate your way to Product Market Fit 4Measure Results and Adjust 5Invest in what works Five Principles 

Ideas Product Data Learn Build Measure Lean Startup Cycle The Lean Startup - Eric Ries 

Ideas Product Data Learn Build Measure Starts with Problem is something customer needs? will they pay for it? can you solve it with current resources? 

Ideas Product Data Learn Build Measure now build an MVP 

Minimum Viable Minimum Viable Product bad product no one wants to use products by company with better resources 

Minimum Viable Product smallest version of your product that allows you to attract users and learn as much as possible from them 

Good MVP 1.enables users achieve their objectives 2.enables us to obtain more data 

Requirement Development QA Release Typical Production Flow some learning very little learning most learning Optimise this so we can invest more time on this and especially this 

Feature A Feature B Feature C Feature A1 Feature B1 Feature C1 Feature B2 Feature C2 Feature A2 release small incrementally instead of do this we may even drop this 

Prioritizing: Pareto Principle 20% 80% Effort Impact 

1.Always optimize for rapid delivery and data collection 2.Only build something you specialize (e.g. algorithm, data, model, business process) 3.A lot of 3rd party tools are free for startups (cloud, analytics, app builder, user interface) Tech: Build vs Buy 

1.Simple Analytics and Dashboard 2.A/B Testing Tools 3.Continuous Integration & Development Tools 4.Platform as a Service 5.Relational Database 6.Low/No-Code Technologies 7.other tech you most comfortable with You Most Probably Need 

1.Self-Managed Kubernetes 2.Microservices 3.Distributed (and/or Non-Relational) Database 4.Kafka / BigQuery / Prometheus 5.Fancy Programming Language or new Javascript framework you don't know about 6.Home-grown Developer Experience Tooling You Less Likely Need 

1.Scrum 2.Kanban 3.Crystal 4.Extreme Programming 5.Dynamic Systems Development Method 6.Feature-Driven Development Agile Developments Methodologies Benefits 1.Flexibility to manage changing priorities 2.Team-centered collaboration 3.Instant visibility with context 4.High quality and faster to market Warning: Don't fall into trap of wasting time discussing unnecessary things e.g "This is not how we should do X according to Y or book Z" 

Rapid Iteration: Continuous Integration Tips: Start from day one! 

We have to pay the debt as early as possible But questions to asks: Does it add value to customers? 1.Cheaper product to afford? 2.Better quality (bug-free & secure)? 3.Faster product delivery? Avoid over-engineering Managing Technical Debt 

Ideas Product Data Learn Build Measure collect as much data 

1.What are you trying to learn? 2.What data you need to collect? 3.What measures as success/failure? Collect maximum amount of validated learning about customers with the least effort 

Vanity vs Metrics Actionable Accessible Auditable 

Stage Metrics Acquisition New Visitors Activation Registration Retention Multiple Visits Referral Referral Code Activated Revenue Transactions Pirate Metrics 

Task 1 As Guest, I want to register using email address & password Task 2 As Data Analyst, I want to measure number of registration by email address & password Task 3 As Guest, I want to register using Google sign-on Task 4: As Data Analyst, I want to measure number of registration by Google SSO Prioritize into backlog Sprint 1 Sprint 2 build the feature implement tracker build dashboard 

Split Testing Validate hypothesis using A/B Testing 50% of users access A, 50% users access B 

Pivot Sometimes your initial hypothesis doesn't work. Course correct steps to test new fundamental hypothesis Study Case: 1.Programming compiler to Operating System 2.Online Games to Image Sharing 3.Ticket Aggregator to Online Travel Agent Tips Engineers: Don't take it personally. 

Ideas Product Data Learn Build Measure There are so much more Build Faster Learn Faster split test customer development five whys product owner falsifiable hypothesis cross-functional teams customer archetype smoke tests unit tests usability testing continuous integration incremental deployments free & open source tools cloud computing refactoring scalability Measure Faster analytics real time monitoring Measure Faster funnel analysis cohort analysis net promoter score 

Conclusion 1 Admit that you don't always have answers 2 Find fastest path to customer experience 3Don't over-engineer. Fail fast; fail early 4Don't get lost in features 5Priorities and plans will always change 

References