• SoNware engineers increasingly choosing startups • Innova2on con2nues to drive majority of growth Today’s talk is mostly a reflec2on on growing gilt from 0 to 1000+ people
to ~2000 • I think of it as having netscape in high school • Only 29% of millenials "expect to work regular office hours” • Development and work/life balance are more important today • Technology inside companies has fla_ened communica2ons • Want flexibility -‐ when to work, where to work from, and how to deliver work • Greater focus on results (vs. tenure for example)
• Inspired most directly by Neclix (culture) and amazon (2 pizza teams and fitness func2ons) • Can we pivot @ gilt to empower small teams to autonomously innovate and deliver value?
essen2ally have one team working on one idea • Over 2me, end up feeling like you can’t get anything done • Unnamed company: “Current backlog is 18 months to get an idea priori2zed”
point, at least at Gilt, we realized we: • Spent lots of 2me talking about ideas that never got done • Did not feel great about the ideas we were building • Felt slow – too hard to gain consensus on approach These were the problems we set about addressing. It hit us hard in our 4th year and roughly ~800 people
an idea, but innova2on is of course having the idea and doing everything else – the R&D, the tes2ng and re-‐designing to develop a new product/service/process, and then successfully gemng it out to the market or pumng it into use, so that it’s not just an idea but an applied result. h_p://original-‐ventures.com/2013/03/13/ideas-‐are-‐ cheap/
what's the difference between the Chicken and the Pig? • Answer: The Chicken is involved, but the Pig is commi_ed! “Pigs” should choose what we build h_ps://en.wikipedia.org/wiki/The_Chicken_and_the_Pig
search • KPIs: • # of users acquired in past 30 days via non branded organic search • Session revenue in past 30 days from users acquired via organic search
“I think we need to make the register bu_on blink.” Team says: “Great idea – our KPI is x and we think our current idea is going to move our KPI towards its target faster.” • Decision making: The team working on the KPI independently selects its ideas to execute on. They are now autonomous and measured only by the KPI
KPIs rolls up to the overall strategy • One common ques2on: “Who does the sh*t work?” • As much as possible, choose KPIs that are independent; where two KPIs conflict, ensure the teams are communica2ng
ini2a2ves • Ini2a2ve is a project we intend to work on for the foreseeable future • Future defined as 3-‐12 months • Evaluated once every 3 months • Each team works on exactly 1 ini2a2ve
“strategic” KPIs • Goal is a single “strategic” KPI where possible; some2mes need a second • Teams may create their own “tac2cal” KPIs to help measure progress on individual projects – tend to be short-‐lived.
product visionary, coder, presenter, visual designer, business thinker, organizer, mo2vator, quality manager, coach, analyst, architect, writer, visual designer and ux specialist • Individuals fulfill 1 or more ingredients • Ideal team has every ingredient
• Review KPI and progress towards targets • Discuss ongoing mul2 variant tests and outcomes • Story: What’s the elevator pitch for next few weeks Works great if the the KPIs are solid and responsive. Hard if KPIs are not great. h_p://tech.gilt.com/post/44792645000/scaling-‐agile-‐at-‐ gilt-‐with-‐teams-‐ingredients [from John Quinn, ex Digg CTO and current SVP engineering at gilt]
make project decisions • To build something that ma_ers • To impact business • To develop a strong sense of well being • To work with great people (great = be_er than me)
It is up to YOU to fill it up with the awesome stuff we're working on and doing here. What you post there is auto syndicated to our gil_ech twi_er account, which you can post to directly. Password=***** Things to Consider • It's oNen a good idea to ask for a peer review before pos2ng to double check technical accuracy, grammar etc. We've also got a team of wordsmithers (see appendix) that have signed up to help you out. • If you've got content that you're not sure about e.g. talking about a feature that's not live or a subject that you're not sure is suitable, ask your manager.
budget • Enter your proposed (or incurred) expense on the next sheet. Your expense should be related to team building e.g. organizing a team building breakfast or a dinner to celebrate a big release. • If the total remaining is now nega2ve, remove your expense. You'll need to wait un2l next month when the budget is refreshed. (the 1st of the month) • If the total remaining is posi2ve, you're approved! • If the budget is exhausted, please add a $0 entry for the thing you would have liked to do. S2ll early – but really what happens is you find yourself always considering if decentraliza2on would be be_er. It oNen turns out to be.
• Focus on providing con2nuous feedback • “Lots of small errors” Ul2mately reinforces having great internal tech – and that by itself leads to higher produc2vity, well being and and happiness
didn’t talk about innova2on today explicitly • But implicitly everything here is about enabling innova2on Nobody has figured out how to create innova2on But we do know what kills innova2on Empowerment and autonomy allow for innova2on to occur