Startup Operating Systems
- 1. Operations Systems for Startups Dean Haritos dean.haritos@soluti8n.com @deanharitos1
- 2. CEO VC COO CEO 1991 2003 1 2 23 34 49 53 CUMULATIVE # of COMPANIES I’m an operations guy mentor director electrical engineer 1995 1997 1999 2009 2014 2016 CEO 2020
- 3. What is Operations? How a company organizes and delivers work
- 4. And we’re talking about Operations Systems … Systems define a process 1. Purpose 2. Inputs 3. Process steps 4. Outputs 5. Quality metrics Once an operations system is well-defined, all work going through that system results in the same quality of output. It becomes predictable and does not need to be revisited.
- 5. Motivation Startups fail for all kinds of reasons ______ Ignoring operations should not be one of them
- 6. PROGRESS T I M E Timing: When today’s talk applies Don’t worry about operations. Invent your thing. Time to ramp up. Focus on operations or fail.
- 7. PROGRESS T I M E IDEAL REALITY Motivation
- 8. PROGRESS T I M E WASTE PREDICT + PLAN NOTICE + RECOVER Deliberately or not, systems develop
- 9. PROGRESS T I M E Would you start writing software without defining the architecture? Be deliberate about your systems The way you set up your company is the way it will run
- 10. Breakout Exercise Choose a department or practice from the right, and define a simple, useful process as follows: 1. Purpose 2. Inputs 3. Process steps 4. Outputs 5. Quality metrics Practices Internal communications Meetings Knowledge management Departments Product development Marketing Sales Manufacturing Customer Support Quality Finance People
- 11. Operations Guy Principles • Treat everyone like a customer • Build systems before teams • Manage by exception • Solve every problem once only • Use trim tabs
- 12. Treat Everyone like a Customer • Intention matters • Operations Systems should not bend employees to the CEO’s will • Your job is to solve problems and create opportunities • For the company • For the people who work there • Treat internal customers and external customers the same • Create value, not bureaucracy • Aim for lightweight systems that accomplish your goals (trim tabs) • Banish the term “change management” from your vocabulary • A new Operations System should be like a highly desirable product • It should be welcomed – not dreaded! – by the recipients • How you sell the system is as important as the system itself • Listening to your customer is the most important part of sales As an aside… many startups originate from Operations Systems
- 13. Operations Guy Principles • Treat everyone like a customer • Build systems before teams • Manage by exception • Solve every problem once only • Use trim tabs
- 14. Build the Operations System First • Respect people • Don’t hire people to run an inefficient process • If you hire teams before creating systems… • You’re going to need to get rid of people • You can’t treat them like customers The right system:human ratio
- 15. Build Intelligent Operations Systems • Be capital efficient • Automate everything possible • Develop tools to invoke human GPUs exactly when needed • Then hire exactly who you need • Practice continuous learning and improvement • Define and conduct experiments • Improve the system to reduce headcount needs to scale • Plan for one initial cycle and one improvement cycle • Then just tweak if and as needed Plan Predict ExecuteLearn Improve 2x
- 16. Define how Work gets Done • Templates • Project planning • Work management
- 17. Define Processes via Fill-in-the-Blank Templates • Fill out the same information every time the process runs • Familiar, trusted format reduces team overhead in evaluating data • Continuously improving the template upgrades every future instantiation Excerpt from NPI Process Template
- 18. Project Planning and Work Management Waterfall is necessary for hardware … but you should also use agile tools (JIRA) Gantt Chart Kanban
- 19. Measure what Matters • SaaS Metrics • Software Operations Metrics
- 20. SaaS Metrics
- 21. Software Operations Metrics Development • Leadtime • Cycle time • Sprint velocity • Open/close rates Production • Mean time between failures • Mean time to recover • Application crash rate Security • Endpoint incidents • Mean time to repair Source: https://techbeacon.com/app-dev-testing/9-metrics-can-make-difference-todays-software-development-teams
- 22. Define what Work to Do • Plan of Record • Product Roadmap • Key Enablers and Platforms
- 23. Italics Blue Green Red = Projected = Done = Win = Loss/miss K E Y 2020 Thematic Goal: Launch Product to Market 2 Quarters Ago Last Quarter This Quarter Next Quarter 2 Quarters Ahead Milestones Month 1 Prototype available Pilot available Receive UL certification Month 2 Pass internal testing Month 3 Alpha customer buy-off Receive UL certification OKRs Objective Raise money to finance our new life-changing product Raise enough money to filet our competition Key Result 1 Pre-money = $2M ß 0.6 Key Result 2 Raise = $750k ß 0.7 Key Result 3 Sales up 10% ß 0.9 Customers Trellis, Gumbo, YKX, Bilge, Karp Trellis, YKX, TunaTech, ZipCo, Funny.com 20 new customers 75 repeat customers 4 lost customers 30 new customers 80 repeat customers 5 lost customers 40 new customers 90 repeat customers 6 lost customers Headcount (EOQ) 3 3 vs. target of 4 5 6 Cash (EOQ) $720k $480k $390k $360k Burn Rate (avg/month) $80k $80k $30k $10k Cash Cliff End 4Q20 Mid 3Q21 3Q23 Funding Raised $500k on $1.2M pre from angels: ShortArms, SmallHands, ShallowPockets - Raise $1M debt financing from ShortArms, BigHands, DeepPockets < Company Name - Confidential > Best Practices: Choose milestones you expect to hit with 70% confidence. Choose key results you expect to hit with 50% confidence. Score past key results 0-1; 0.6-0.7 is your target. Miscellany: EOQ = end of quarter Cash Cliff is the month you run out of cash BHAG definition Learn about OKRs and how Google uses them Plan of Record
- 24. Product and Enabler Roadmap
- 25. … and then you can Hire • Right skill sets • Right cultural fit • Right number of people
- 26. Operations Guy Principles • Treat everyone like a customer • Build systems before teams • Manage by exception • Solve every problem once only • Use trim tabs
- 27. Manage by Exception • Set up red / yellow / green indicators • A moment spent talking about something on track is probably wasted • Pay attention to what’s not working • Trust but verify • Don’t let your system lie to you • You still have to look under the hood (“Management by Walking Around”) • If you have a hole in the system, and you blindly trust the system, you’re dead • Hold serious reviews of key features, milestones, etc.
- 28. Operations Guy Principles • Treat everyone like a customer • Build systems before teams • Manage by exception • Solve every problem once only • Use trim tabs
- 29. Problems Happen • Problems will arise • Count on it, don’t dread finding them • Every problem discovered is an opportunity to improve the company • How you treat the situation is a cultural issue • Nobody wakes up and says, “I’m going to intentionally screw up today” • My approach is to defuse the tense emotional charge with a joke: • “OK, let’s go to step 1 in a crisis situation: assign blame” • “Oh crap, everybody panic!” … and then to identify the root cause and prevent even a single recurrence
- 30. People Management • If you uncover an employee performance problem, address it • Startups do not have time to keep the wrong person in any role • Your HR process should take over from there; not today’s topic • Bottom line: solve the problem completely
- 31. Operations Guy Principles • Treat everyone like a customer • Build systems before teams • Manage by exception • Solve every problem once only • Use trim tabs
- 32. Trim Tabs in Business Study existing dynamics Find inflection point Apply minimal pressure for desired outcome Benefits Stabilize the system in new position Eliminate effort to maintain control Neutral Position Light Tug at 1 1 Use Trim Tabs
- 33. Trim Tab: Building Manufacturing Engineering at Finisar Inflection Dynamics Stability No Effort Fiber optics industry weak in manufacturing Disk drive industry imploding Volume manufacturing impossible to teach Strong engineers can learn optics Hire very best disk drive engineers Create fiber optics course Endless supply of engineers Core expertise in volume manufacturing Finisar Mfg Engineering on par with R&D Disk drive engineers excited to learn new field People recruit their colleagues to join Pressure
- 36. Effective meetings • Have drama • Fit within an architecture of magnifying glasses and meetings • Death by Meeting (recommend reading all Lencioni books) • Examples: Glass boards vs. white boards Throw away all the two-ended markers
- 37. Airing of Grievances • This should be the one all-hands meeting that nobody is allowed to miss • Participate in person if at all possible (I used to not let anyone call into the meeting because it changes the energy dynamic in the room) • Celebrate wins • Bring up and analyze losses • Explore / communicate company changes and get buy-in • Airing of Grievances • The name makes it fun and reduces barriers to speaking up • Field random questions and complaints • Keep it lighthearted • Focus on understanding the grievance (in here is assessing validity, but it will come out naturally from the whole team if the meeting is working properly… “who else is having this problem?” “what do we all think about this?”…) and thinking about what the solution might be • Eliminates FUD (major drag on productivity)
- 38. Operations Project in JIRA / Suggestion Box • Unfrustrated employees are more productive Concept of “fire and forget it” – offloading the frustration into JIRA and knowing it will get addressed removes the problem from human RAM so the person can focus on his/her job Can always go back and visit to see how the solution is coming along, and be part of the solution if desired • % mind share on what they were hired to do approaches 100%