ProductSF 2014

Photo by Ritu Raj
Photo by Ritu Raj

Last week I attended the 2014 edition of Building Better Products, an invitation only conference hosted by Ty Ahmad-Taylor of Samsung Electronics and Josh Elman of Greylock Partners. As with last year’s conference, it was an agenda packed with interesting speakers that came to share their perspective on the role of a Product Manager at a tech company, sharing lessons learned and what it takes to succeed. The format was updated to include a mix of pechakucha-style lightning talks interspersed with longer presentations and panel discussions.

Because of a morning SFO drop-off run, I missed the “palette-cleanser” talk that kicked things off for the day (a history of Treasure Island) and Julie Zhuo‘s talk on how to work with designers. I’m sorry to have missed Julie’s talk but you can read her many excellent talks about the design process on her Medium page.

I took notes on some of the presentations and here were my favorites.

Mina Radhakrishnan shared her experiences as one of the first Product Managers at Uber during their first international expansion and the growing pains the company went through to develop a process which now lets them spin up multiple locations each week. We learned of design challenges as the nomenclature for hired cars vary from city to city. An SUV in London is a “big car” while a private car is a “black car” but tabs for “Big” and “Black” don’t really work so other solutions were needed.

A template that might work in one country falls apart in another so you need to, “design for flexibility” and use templates that can be adopted to fit local needs. Because of the need for speed and scale, these templates need to be designed in a way that local operations teams can unleash their “operational creativity” without having to check in with headquarters.

What struck me most was the intense drive and ambition of Uber CEO, Travis Kalanick, enabled the team to roll over technical and legal challenges that would have sidetracked most companies. In the drive to launch Uber in Paris in time for Le Web, they slammed a product together that had to be re-built from the ground up post-launch. It is rare to find a start-up that would knowingly build something they know they are going to tear apart and build again and support for that way of working has to come from the top.

ubercade

While talking about Uber’s various publicity stunts (ubercade, uber ice cream, uberkittens) Mina reminded us that Uber is not strictly a transportation company and that their mission, “to get you where you want to go” should be interpreted broadly.

Vince Maniago from Mint gave an abbreviated version of the talk embedded below to explain specific actions he took to improve sign-ups. As a financial service, his challenge was to get people to connect their banking accounts with the mint.com aggregation service. By putting his name on follow-up emails to anyone who didn’t complete this action, he was able to significantly improve completion rates and get people over the trust factor.

Some other tips Vince shared included use of Amazon’s Mechanical Turk service for user-testing and Mental Notes cards (note: sold out) as inspiration for the types of message variations you can test.

David Hahn, who was at LinkedIn during the early days, spoke about the freemium business model upon which that service is built and the evolution of their landing pages. The page has gone through many iterations but each time they have tried to add something new, they have found that the obsessive simplification of the erroneous was the best contribution to their conversion rates. “The best optimizations are getting rid of things.”

Wook Chung works with brand advertisers for Samsung and shared humorous anecdotes about working with Madison Avenue and the importance of continually checking in with your customer whose needs are like Heisenberg’s Uncertainty Principle. Advertising trends are continually in motion so freezing requirements and then going off to build them will, by definition, obsolete anything you launch. Better to involve your customer at each stage and get them invested in your solution and inform your development.

A series of short talks followed. Ken Norton (Google Ventures) on using Don Lowry’s True Colors rubric to build teams and modulate communication. Ian Spalter, UX Lead at YouTube, gave a rousing comparison of UX/Product archetypes as Game of Thrones characters and reminded all of us to, “respect the craft” when giving feedback.

During the panel discussion on Product Development in the Enterprise Sol Lipman, who was CPO at TomFoolery, shared lessons learned as a start-up trying to change enterprise software (“we made consumer-grade products for the enterprise”). One of their early lessons (and we’ve heard this before) is that you need to focus on utilization because, “just because someone is going to pay you doesn’t mean their people are going to use your product.”

One other tidbit was Sol’s singular vision that the all products for the enterprise should be built “sideways” to allow collaboration across traditional networks. Do not build your product around a company domain and allow for self-defined networks which will be how workgroups form in the future (think Dropbox). Email, which is usually dominated by company domains, is not a useful unique identifier as it locks each customer into their silo. On the other hand, a cell phone number, which everyone has and usually carries from job to job, allows for infinite connections which will persist and retain their viral effectiveness.

tomfoolery hats

Ellen Chisa from Kickstarter shared the eight-month journey to a new start page. Reminding us all that Kickstarter is primarily a platform for artists so the founders were concerned less with a metrics-driven design process and more with the embodiment of a vision. If forgot the name of the poet she quoted but, to paraphrase, if Kickstarter was going to set out to build a ship, they would not instruct them the details of woodworking, they would, “teach men to yearn for the sea.”

In the end, there were 11 major designs for the front page with hundreds of iterations in between until they found a design they felt worked. But all was not lost, many of the ideas hashed out made their way into other areas of the product. Perhaps less efficient, the Kickstarter way was more holistic.

Shiva Rajaraman, from YouTube, gave a great talk, “Product Management: From Meh to Awesome” that was not about use cases or tips and tricks but about approach. A Product Manager is the one that has the spotlight. No product is the result of a single person’s output and the PM should take pains to shine attention to all those that work on the product. And with a community-driven product such as YouTube it is especially important to point out what’s awesome about your community.

Check out the Likes/Dislikes on the Futurama Neutral Response clip below. Finding out that there is a community of people that obsessively maintain this delicate balance, “made his week.”

futurama neutralness

Thinking of YouTube as a platform transformed the way they have approached music companies and brands. Music fingerprinting and content ID was originally developed to go after copyright violations but when flipped around from a negative to a positive, it helps musicians understand how and where their music is being played. Harem Shake got such tremendous exposure through the fan videos that when it debuted on the charts, it was #1.

Shiva tells the story of Molly Kate Kestner who recorded her self-composed song on a broken iPhone that, within weeks becomes a top single on the iTunes charts.  An open platform  allows these stories to write themselves and a good Product Manager will look for them and hold them up for others to celebrate.

Thank you Josh & Ty for organizing a great day of talks, picking a great venue, and, as always, keeping everyone on schedule.

Further Reading:

The Art of Product Management

View from Product SF

Many start up Product Managers are lone wolf types. They work alone, part of a team, but the only one in that team doing what they do. Shuttling between engineering, design, sales, marketing, management and other constituents, they are the glue that ties the team together, the universal joint making sure all the momentum in one group transmits cleanly to the next so all the pieces move in unison and things get done.

Because they work alone, Product Managers have developed their own set of tools that work for them and their company. Tips and tricks on how to organize teams and set priorities are passed around like folklore on blogs and Quora posts but it’s rare to find an event that gets a room full of PMs sharing what works in real-time.

Perhaps it’s a sign of a maturing industry legitimizing a role but over the past several weeks, I have had the good fortune to attend three (!) events specifically for the modern Product Managers.

Each of these events were great in their own right but it was this past Friday’s How to Build Great Products hosted by Ty Ahmad-Taylor of Samsung and Josh Elman of Greylock Partners really knocked it out of the park.

 

What I Learned

Many PMs fell into their role before they realized that they were doing it. Especially at small start ups on a hyper growth trajectory, the PM is the first to jump in to rally and organize and bring teams together and ends up creating their job.

PMs need to have a singular focus on what needs to get done to make their product successful. This means sometimes going against corporate objectives. Hunter Walk, when he was a PM at YouTube, put priority on supporting Facebook login over Google’s own competitive Open Social login initiative. You need to look at the market as a consumer and make the choice based on what’s right for the customer, even if it sometimes goes against larger corporate initiatives.

PMs need to be entrepreneurs. Faced with a limited amount of resources, you need do whatever it takes to get things done. This often means you need to think creatively. Hunter also mentioned “learning to draft ascending ecosystems” – take an objective look at the market and if there is a way for you to bundle or integrate your product into another product on the rise. By focusing on getting the YouTube icon added to the default iPhone deck, all other conversations with carriers turned from YouTube chasing them to the carriers calling YouTube. Picking the ascending leader early was a bet that paid off.

Trust Your Strategy. Allen Blue, the co-founder of LinkedIn, described the first year of nominal growth (2,100 users after one week, 80,000 after 7 months). They reserved a bottle of whiskey for when they hit 1 million users but ended up drinking at 500,000.

PMs are problem solvers, give them meaty problems to chew on. Like Allen, many PMs come from varied backgrounds. Allen came from the theatre. Because of their non-traditional training (Hunter Walk too, was one of the few PMs at Google without a CS degree), a PMs can be the source of a unique perspective and often solution. One useful technique to use when drilling down to the essence of a problem is by using the 5 Hows which is a derivative of the 5 Whys, designed to lead you to a solution, not causality.

Srikanth Rajagopalan on building the Chrome Browser

“Speed is a feature.” This is often ignored when prioritizing tasks. Do not forget that simplicity, and it’s cousin, speed, are silent feature requests that should always be potentials for any roadmap discussion.

Build for the average user but do not ignore your super fans by hiding power user features. Chrome has a number of geeky features (Keyboard shortcuts that one could only appreciate if you, “grew up with vi”) but they are tucked away and do not get in the way of the basic user’s core experience. Deciding how to add an advanced feature is difficult. Putting it into Settings and letting the user choose is lazy. Did you know  Chrome has a detailed Site Info box?

Chrome Site Info

There was a panel discussion about user growth where I unfortunately had to step out for a bit but I did catch Elliot Shmukler talk about the power of shared metrics. During his days at LinkedIn, he pulled together the most important metrics into an information radiator so that everyone gets a realtime view of these metrics with just a glance. But one must not forget to focus on the right metrics. Too often a dashboard is swamped with metrics where it’s too easy to drown in data. Pick 3-5 metrics and use those to test intuition. Be careful which metrics you chose because those will be the ones you optimize for.

A funny anecdote Elliot shared was coming in one morning to find, “Belgium on fire.” User acquisition spiked overnight because an engineer, after looking at these shared metrics, fiddled with some of the contact importing settings for users in that country to see if it would move the needle. It did. That is the power of shared metrics.

Dan Olsen gave an entertaining talk about his time at Friendster during their meltdown and subsequent loss to MySpace. See slide 21 which shows how he came up with the concept of the viral loop (no one was doing this kind of stuff then) and how different points in that loop have different amplification values.

Despite succeeding in rolling out several new features that helped in user growth, we all know the story about poor database optimization that ultimately ruined the experience on Friendster so that the more users joined, the slower the site became. Dan’s learnings were that, for social networks, a large user base is a feature. People used MySpace because that was where the majority of their friends were which lead to, um, a majority of their friends being on MySpace.

 

Fred Sigman, an Art Historian & Photographer, gave an thoughtful talk about the influence of nearby atomic bomb testing on the neon signs of early Las Vegas. As a post-lunch talk, professor Sigman’s talk was perfect as the, “palette cleanser” talk before we got to the. . .

Ignite talks. These are lightning round, Pechakucha-style presentations where the slides self-advance. Highlights for me were Ken Norton’s How (not) to work with Engineers and Ian McAllister’s talk about the User-Driven Development which I’ve written about before.

Ian Mcallister

Joe Zadeh from AirBnB spoke about the hiring process at his company. Forgive the fuzzy resolution but it’s the nut of his talk which explained that all PMs that are interviewed are invited in to present on a topic given to candidates in advance. This allows them to evaluate all the candidate along a baseline for a number of things simultaneously,

  • Can they communicate effectively?
  • What’s their design skill like? Do they understand spacing, fonts, color, use of graphics?
  • Are they nervous presenting? How do they respond to questions, pressure?
  • Did they have a fresh approach? Are they an original thinker?

If the interview is for a Product Manager, they may make it about a specific problem they are trying to solve. When they were looking into adding a payments processor in South America, they asked prospects to speak that topic. By the end of the interviews, they had not only a good sense of who knew what, they also got six different approaches to integrating payments processors.

Joe Zadeh

James Buckhouse from Twitter gave an great talk about storytelling. @buckhouse came from Dreamworks and tells stories for twitter and shared his formula for how to tell a good story. It’s all about the journey to a transformation. He writes about writing on Medium at Design Story.

James Buckhouse on Story Telling

The final talk of the day was with Tom Conrad at Pandora. No slides, just a simple discussion between Josh Elman and Tom about the early days of Pandora and it’s inspiration (“I loved to recommend music to people. Building a service was a lot easier than inviting people to my dorm room.”) but also included a varient of the “let’s go shopping” style of roadmap prioritization.

  1. Every 90 days, wipe your roadmap clean.
  2. Ask everyone in the company what things the company would be crazy not to do. This includes projects that used to be on the roadmap but are now off it.
  3. Summarize each idea on one slide. Work with engineering to come up with a thumbnail estimate for how many days it’ll take for each project. If a project will take one weeks’ development by two developers, that’s a $10 project.
  4. Print out each slide and stick them up on a wall.
  5. Sit with Engineering and decide how many developer man-days you have for the next 90 days. If it’s 60 working days and you have 2 developers, and you need to reserve 10 days for maintenance, that means you have 110 developer man-days. This would equal $120.
  6. Invite a cross-representative group of people to a product prioritization meeting. At Pandora, this included someone from HR. Give each person a different colored post-it note with equal amounts that they can allocate towards projects. If there are 5 people in the room, then that’s $24 each.
  7. It will be a long meeting but, by the end, you’ll clearly see which projects are “fully funded” and those will be the ones that you take on for the next 90 days.

Everyone leaving that room will have a clear view of what engineering is going to be working on better than any Roadmap document could tell them. There will be no need to brief people on what each project is because they will have heard the description of each project in that meeting.

I hope you found this post useful. I’m gathering these tips and tricks into a collection. If there are some that you feel should be included, let me know.