Task Labor: Every Worker a Small Business

Fifty pounds of grass-fed beef sat thawing, forty miles away

Some friends and I had ordered a quarter of a cow and it was delivered to their house on a Friday afternoon. “You’d better come pick up this meat,” my friend said, “it won’t fit in our freezer.” Driving for hours was not how I’d planned to spend Friday night.

TaskRabbit to the rescue.

A TaskRabbiter named Sean drove to San Jose, picked up the meat, and drove it back to my house in Walnut Creek. It took him over 3 hours. I paid him $35.

Task-labor is one of the most convenient and concerning new developments in modern living. Every couple of days a new on-demand labor service launches: On-demand dog walkers with Swifto, mobile massages with Zeelon demand dry-cleaning with Dashlocker. There are services for on-demand mechanicshouse cleanerstutors, doctorssnow plows, and many more

In one sense, I’m excited about the convenience that these services offer. I was thrilled that it was Sean driving for three hours to pick up the meat, and not me. That said, I think my arrangement with Sean represents one of the fundamental problems with task-labor: no protection for the workers. To work as a laborer in many of these services, you are an independent contractor and not an employee. That means:

  • Labor-side profit risk—Unless Sean batched several deliveries, my task was a financial loss for him (but not for TaskRabbit).
  • No recompense—When Sean discovered that the delivery would be unprofitable, he still had to complete the task (for no extra money) or risk getting kicked off the platform.
  • No guarantee of work—There is no job security in task-labor. In a typical corporation, the company holds reserves to smooth over the dips in demand. In task-labor, workers directly feel the ebb and flow of demand. [1] Actually, they probably feel the ebb more than the flow: in times of high demand, 1. there are only so many tasks a single worker can complete in a day, and 2. service providers hire more laborers to reduce wait time for their customers.
  • No insurance protection—Sean drove his own car and, beyond that, he was directly liable for any damages that may have occurred as a result of his services.
  • No benefits—Task-laborers must purchase their own health insurance. Individually-acquired pre-Obamacare health insurance for a family of four is $1700 a month in the Bay Area. I.e. it’s unaffordable to almost all task-laborers.

The service platforms are very careful in the way that they pitch to laborers. You get to be your own boss. You’re a small business owner. You’re venturing out as a “micro-entrepreneur.” But hidden beneath these labels is simply the fact that costs normally covered by an employer are shifted to labor.

An employee is offered predictable income, health insurance, and employees suffer no personal loss if the company fails to turn a profit. But in a task-labor system, you are not guaranteed a profitable (much less livable) wage.
 
Task-labor is the realization of a capitalist ideal: own the means of production and let labor fend for themselves.

Certainly, it is not a new phenomenon for people to provide short-term labor for individual tasks. When I go to Home Depot there are always a dozen men waiting outside looking for work (although no one labels these men as “micro-entrepreneurs”).

The techno-solutionist in me wonders, “Why are these men still here? Why aren’t they all on TaskRabbit by now?”

It turns out that sharing-economy task-labor is still unavailable to the poor. Being able to sign up for one of these services requires that you 1. have a smartphone with a data plan, 2. pass a background check, and 3. possess good English skills. 

So then where is the task-labor force coming from?

If we want to assign directionality to this phenomenon, I’d say people are moving “down” from wage paying jobs rather than “up” from unemployment. Task-labor is an increasingly populated rung between wage-labor and unemployment.

Proponents will argue that this is one of the benefits of task-labor: if these people did not have task-labor, then they wouldn’t have any work at all. Task-labor is supporting the bottom end of the economy and, without it, people would be worse off.

While this is probably true, this argument sounds like spin on an incidental dynamic. These services are for-profit, not the C.C.C.  and labor is a cost. For example, when driverless cars come to market, the on-demand delivery companies will fire the delivery people and keep only the employees needed to support logistics at the endpoints.

Because the field is so new, these laborers are relatively unsophisticated and do not (yet) realize the costs and risks that have been transferred to them. This poses a problem for the service providers: on the one hand, paying for increased benefits will drive up costs and squeeze margins. On the other hand, no ruling class wants to deal with uprisings from a disgruntled labor force. 

Task labor service providers want to avoid any cost increases that are out of their control (particularly minimum wage legislation and unionization). I expect they will act first and institute some form of self-regulation which will provide services to help laborers.

Here are a few of my ideas on how these services can create a better environment for task-laborers:  

  • Profit-aware tools—Enable laborers to make decisions about the profitability of a given task. Help them calculate the most profitable tasks given location, traffic, gas, and expected completion time.
  • Suggest logistical task combinations—Enable laborers to combine activities (possibly across services: e.g. allow someone to make deliveries for both Postmates and Instacart concurrently).
  • Micro-benefits—If we’re enlisting someone for micro-labor, maybe they should be eligible for micro-benefits. Someone who works for several services could, in aggregate, receive the equivalent of full benefits.
  • Liability insurance— I think a good rule of thumb here is, if a worker is operating in your name, under your brand, and you share some of the profits, then you should share some of the liability if something goes wrong
It is undeniable that task-labor services are convenient if you can afford it, but I’m conflicted about the societal impact. On the one hand, if you think that minimum wage is unnecessarily high then you’ll view TaskRabbit as a good thing because it corrects legislative market meddling. On the other hand, we’re labeling poor people as micro-entrepreneurs to avoid giving them benefits. The laborer isn’t a person – he’s a small business.
 
It’s important to note that some of the on-demand-labor services are almost certainly a net positive. For instance, HomeJoy has a reputation for being beneficial to both labor and customers by paying more than minimum wage but still charging less than traditional competitors. Exec dealt with the problem of labor-profit-risk by charging per-hour fees for it’s assistants. However, it seems that the demand wasn’t high enough for this service, presumably because customers weren’t willing to take the risk of variable pricing. Exec has since switched to housecleaning.  
 
Despite the shortcomings many of these services have, the economics seem to be working for some laborers: Sean is still active on TaskRabbit. I know because I asked him to make another delivery for me. This time I bought a $900 Herman-Miller chair from a Silicon Valley fire-sale. “The last thing I’m going to do is drive all the way to Palo Alto just to pick up a chair,” I muttered. TaskRabbit to the rescue.
 
See also:
Posted in programming | 1 Comment

Growth Hackers Conference 2013 Notes

 

Yesterday I attended the 2013 Growth Hackers Conference at the Computer History Museum in Mountain View. Below are my notes, typos and all. There are more than a few gems to be found below. If you’re short of time, I particularly recommend the first and last talks.

James Currier – Ooga Labs – Your Growth Fingerprint

  • Grew 10 companies to 10 million users or more
  • His perspective is from a lot of failure, and eventually you’ll have a success
  • There isn’t just one way to grow, there isn’t just one trick
  • Growth is an approach, a philosophy
  • Each company is entirely unique
  • Your fingerprint:
    • Number of Users Today
    • Language-market fit
      • vs. Product-market fit
      • vs. Founder-product Fit
      • The market has to understand your product through the language you use
      • Language is the foundation of growth
    • Growth type
      • it’s not fixed – it changes over time
    • Once you understand the fingerprint, THEN you can talk about tactics
  • Users
    • four buckets
    • 0 – 500
    • 500 – 10k
    • 10k – 1M
    • 1M+
    • underlying blog posts and speeches people are talking about different stages. they may not point it out, but you need to see it
  • Language: what IS it to the user?
    • the user has a full life
    • your product is one tiny part of their life. They ask “what is it to me? where does it fit in”
      • Product Strategy Product marketing Positioning
      • Language and psychology is a faster way to arrive at something
    • What is the psychology – without some sort of insight into your user’s psychology, you’re unlikely to build a great company
    • over the next 5 years you’re going to see consumer psychology applied to enterprise. its a huge opportunity and the psychology applies both ways
  • Language examples
    • store your photos vs. share your photos
      • once we thought of it that way, new language and new features emerged
      • its a language change for the users and for you
    • Find a Date vs. Help people find a date
      • language change changed how people related to the site
      • someone who is married could sign up and help other people
      • that language shift allowed it to get viral and changed our view of the website
  • Language changes your thinking about the product
  • Type of growth you need today
    • People think of growth as new users – “topline growth”
    • Next : activated growth
      • mobile devices, we see this more
    • Next : retained growth
    • Next : monetized growth
      • The reason they installed the app is different than the reasons that are needed to monetize
    • Focus on retained or monetized first before you focus on new user growth. If they’re not sticking, don’t bother with topline growth
  • Now you can think about “tactics”
    • Absorb the tactics where you are
  • There are not that many ways to think about monetization growth
    • 18 revenue tactics
      • type addition
      • placement
      • … 16 more I didn’t get
    • 6 retention growth
      • time
      • threshold
      • availability
      • human action
      • promise
      • deal
    • Activation growth is generally FUE – first user experience
      • 1/3 building game, 1/3 building instrumentation for metrics, 1/3 messing with FUE, redoing and redoing and redoing the FUE
      • its the first time they’re interested in the context, the language, and why they should come back
      • i didn’t understand the level of effort. I didn’t understand that 1/3rd of the time would be in this experience:w. the proportion of time required
    • New User Growth
      • WOM – word of mouth
      • Viral – instrumented viral – classic look at flow / drop off / invite rate up. Instrument the growth. People responding to viral emails, Facebook platform. Recently it’s harder.
      • Paid
      • Within each you have a limited number of channels
      • WOM is resurgent
        • mobile devices put your app where the people are talking face to face
        • growth is happening in between platforms these days
        • there used to be email and SEO and that was it
        • but now we have so many more / email -> android -> twitter.
        • e.g. growing across platforms simultaneously
      • Instrumented Viral
        • activation / entertainment / power up / incentives
      • Be aware: it’s hard to do more than one operationally (e.g. pick one WOM / Viral / Paid)
        • its a human resources issue but also a language design and product design issue
        • it’s hard to switch
        • if you’ve got something that’s working – double down on that
    • Increase all growth approaches & types
      • Reduce page load time
        • Jerry yang said no page can be over 40k in size. yahoo vs. infoseq
      • Reduce cycle time b/t steps & b/t uses
        • evite – cycle time problem. once every 6 months
      • Add more paths/rewards/motivations
        • incentivizing people with money is generally a bad idea, bu
      • Build a culture of growth
        • Simplicity
        • Rigorous simplicity
        • We want to build this beautiful thing for people, but beautiful things can get complicated.
        • If you build 5 features, and people only use 2, throw away the other 3. It’s hard to tell people to rip out features they made
        • You have to educate people when they join the company that that’s what you’re going to do
        • Sustain pain of failure and iterate relentlessly
        • you have to be willing to come to work and have the experiment fail again and again and again
        • you have to educate your team that “we are making progress and this is how”
        • unless you educate your team that this is okay, it gets really wearing and teams break down
        • You must be patient
        • every so often a channel comes around that will 2x 3x 10x your sign ups
        • every now and again massive growth opportunities occur, but they’re not always present
        • you have to lay the ground work
        • when you see the opportunity you have to jump on it
        • It’s not a one time gimmick, it’s a systematic approach, a framework, a philosophy
    • Q & A
      • Enterprise
        • become really good friends with your users
        • beg for the negative feedback – “why would you even use this?”
        • you will have some data on FUE e.g. retention
        • if you send out your team, e.g. head of marketing, they want the process to work. they want to come back and pitch that it worked so they can sleep at night. but you have to educate them that you have to seek negative feedback and that is how we get better
      • Growth Team
        • curiosity
        • don’t get too strongly attached to their own ideas
        • growth team should be more aggressive than the CEO
        • you want the CEO saying “hold on, lets not get that aggressive”

Jared Fliesler – Insights from Growing Square

  • With Facebook
    • It was a wild wild west
    • With tactics you could get 10mm users a day and everyone would use those tactics
  • People tend to think about growth hacking as optimizing the homepage
  • But real growth comes from a product focus
  • Certainly its worthwhile to optimize your sign-up flow
  • Product / Marketing / Partnerships
  • What are changes we can make to the product to open up to new customers?
  • Three themes
    • Solve a big problem
      • If you have a technology first, then you’re trying to grow demand and the product
    • Craft the experience
      • Being delicate, thinking about every detail end to end. Not just from a utility standpoint, but from a
    • Build community
      • Lyft, uber, any company with explosive growth has someone to manage the community community so that people can spread the product
  • Am I solving for something that people actually want? How am I making people feel?
  • LTV
    • Conversion rate customers acquired retention engagement revenue – square X virality
    • we wanted people to care about square
    • crafting an experience
  • Focus on BIG channels and programs
  • This talk was basically “look how great we’re doing at square. btw make your product good too”

Elliot Schmukler – The 3 Steps to User Engagement – Lessons from LinkedIn

  • Speaker
    • VP Product and Growth at Wealthfront
    • Led LinkedIn’s Growth from 20M to 200M
  • It’s not enough to sign up users, you have to engage them too
  • Background
    • 13MM users when he started at Linkedin – now 200MM
    • Re-starting the growth team – 2008 -> 2013
  • The 3 steps to engagement
    • Find you’re aha moment
    • use email
    • engage the already engaged more
  • Finding The Aha-moment = a moment when you’re user truly experiences and “Gets” the value of your product
    • In Linkedin – if you see a profile of someone you know then you instantly understand what Linkedin is about, e.g. my resume on the internet
    • Getting an email from a recruiter on Linkedin
      • This helps you recognize the value of this service
      • “If I have a profile I have opportunities”
      • Even if you hate the recruiter and the opportunity and you never would use it, you still keep your Linkedin open because you see the opportunity and think the next one will be good
    • Aha moment with uber is when you press a button, request a car, it actually comes, and you drive off
    • Find the aha moment by understanding what actions users need to take on your site to become long-term engaged
    • Understand what correlates to later user engagement
    • Examples
      • FB: 10 friends in 7 days
      • Linkedin: 30 connects in 30 days
      • Wealthfront: a 5–10k initial investment
  • Use email
    • You should probably send more emails, but make them highly relevant
    • If it’s mobile, you have push notifications, badges. These are very similar so the rules generally apply
    • You won’t be successful with email unless it is relevant
    • Relevant emails are:
      • From or about someone I care about
      • sent at the right time and frequency
      • filled with interesting content
    • People like getting emails about themselves: “Who’s been viewing your profile”. It plays to my vanity. How popular is my profile?
    • Leverage vanity – the person I care most about is myself
    • Endorsement is a twofer – it’s from someone I care about and it’s about me
    • Think about the frequency and the timing of email.
      • We often think to alert a user right away – this isn’t always the right approach
      • “These are all the people in your network who changed jobs” – but twice a year it was brought together in one email
      • People changing jobs was the most interesting content Linkedin had
  • Engage the already engaged more
    • It’s easier to build on a strength than to improve a weakness
    • It’s easier to get an active user to do more than to get an inactive user to do anything
    • If you want to lift engagement, you should be looking to your active users to see how they got there, but you should also look to see if you can engage them even more
    • Examples
      • Who’s been viewing your profile – You have to have a pretty good profile to even get this email. Inactive user’s don’t get this.
      • If you’re on Linkedin looking at your own profile they’ll ask you questions about yourself which helps you become more engaged
      • “People you may know” – a feature that mainly active users user. But notice what’s going on here – when I connect with someone on this page, then I may bring an inactive user back to the product. E.g. the active users reengage the inactive users
      • Similarly, when you endorse one of your connections, they receive an email so you’re able to reengage a potentially inactive user.
        @eshmu
  • Q & A
    • What do you test in email?
      • length – shorter is better
      • call to action – strong language and button
      • subject line – what works
      • landing page matters too
    • Profile
      • we let people build their profile in Linkedin in bite sized chunks – you don’t have to build the whole thing at one time
    • Growth features
      • how do you measure the annoyance and effect of growth efforts
      • are they working ? attention and activity rates
    • Why does the profile page max at 500+ connections
      • The site isn’t that useful if you have thousands of connections
      • And it has an effect where people work to get 500 connections
    • Are growth strategies different for finance and Linkedin?
      • Certainly – Linkedin was built for virality.

Dan Martell – Retention to test product/market fit

  • Intro
    • Clarity / Flowtown / Spheric
    • @danmartell
    • Started 5 companies – first 2 were failures
  • Often wrong, never in doubt
    • The growth work is a process of being wrong most of the time until you eventually figure it out
    • If you have a SaaS business then the goal is to get new users from trial you need 20% of your users to come back on a monthly basis
    • From: Sean Ellis – 50 new sign-ups per day, 350 per week
    • You want enough new users that you can run experiments against
    • Brant’s customer development survey: ask your users, how would they feel if they could never use your app?
    • Would a percentage of them actually care if your product never existed? You need 40% to say “very disappointed”
    • You want to know this information early – you need to find out what’s not working initially
  • Content marketing
  • For Clarity he hacked podcasts
    • Who are the speakers, notable people in this space
    • I searched for their names in iTunes
    • Since I was in that space, I assumed that I could speak on those podcasts too
    • That script is automated – if someone does an interview, I automatically know
    • People go back to the beginning of podcasts, so we get new customers all the time
  • Press
    • Go find other companies that don’t compete against you in the same space and meet w/ the CEOs and come up with a story to pitch to the press
    • Press gets you traffic
    • Blogs are worth investing in because you can channel your audience
  • Measure
    • What’s the purpose of a business?
    • The purpose of a business is to create and keep a customer
    • Engagement and retention is what helps make something go viral
    • Dan measures everything – he didn’t do this in the beginning and it was an expensive lesson
    • He told a story about how there were no new sign-ups for 3 weeks – an engineer had changed the CSS so it was impossible to sign up
    • You must understand your retention metrics
    • Every product tells a story about what you do – you need to consciously be an editor of that narrative
    • They measure their users by age cohort. Retention will change based on the perceived usage of a product. E.g. every user is tagged with the flow they signed up for (forever!)
    • “How would you describe your company to a colleague or a friend?”
  • Notifications
    • Three types of get your user back to your site:
      • Building something that is so great that they come back on their own
      • Reactivation – having somebody else mention it to them, being reminded via social media
      • Reminding users that you exist
        • Engineers don’t want to bother their users (they don’t want to feel like a spammer)
        • But users want to be reminded – they signed up for your app
        • Users arrived, read the benefit statement, signed up
        • Notifications, emails, phone calls
        • In your sign up you should have a box “sign up for free phone support” – always. If they say yes to that, then ask for their first name and phone number
        • The goal here is just to help them have a better day
      • In those channels you have
        • “Social notifications” – don’t do this please. sending a user a message because their friend signed up
        • “Teaching sells” synthetic – things that you can teach your users. “Teaching sells”
        • Teach your customers how your product works
        • Teach your customers how your other customers have had success. Tell your customers about how other customers use your product.
        • “smile and dial Thursdays”
        • happens every Thursday – a list of your customers that have signed up (goes in timezone order).
        • Don’t tell them you’re the CEO
        • Smile and dial Thursdays is him asking customers “how can we make the product better to suit your needs”
        • I don’t try to sell the on the product, I just try to listen
        • It’s my opportunity to hear how our existing users feel about our product
        • Where I can find where the value is that our product delivers
        • A lot of founders don’t have enough value in the product to help their customers have that “a ha” moment
        • (it also reminds them about your service)
        • The most potent thing I do as a product CEO
      • Two tactics
        • Usertesting.com – I buy user testing and run it against my competitors

        • Watch the videos and see what gains they can use

        • Hired someone on odesk to sign up for my competitors site and created a timeline of all of the notifications and messaging they’ve received from them

        *

        * @danmartell

      • Q & A
        • When is the best time?
        • When they’re awake! Just do it. I have an admin dashboard of users and their timezones.
        • Don’t optimize it at first, just instrument it.
        • How to find a time to build teaching content?
        • I write down when I talk to people
        • If I find myself teaching people over the phone, I write it out. If I find myself teaching the same thing over and over then I produce content based on that.
        • I made an open google doc and I sent it to 25 power users. I told them to add their tips and they’ll be featured
        • Intercom: I invite users of the app to write a blog post, but it’s automated by Klout, Social, area of expertise. This helps me get guest bloggers
        • Enterprise lead time
        • The person who is buying isn’t always the person using.
        • Tax software is only once a year, email newsletter might be once a month – you have to come up with your own metrics
        • Find the users that have had success and look at the click-path / session-stream and look at what their experience was in the product
        • How would you hire growth people
        • Ask for advice first and then if you have people you like then offer to hire them
        • Get to know people
        • Hiring can be like getting married before even going on a date – you have to see if you like each other
        • Seth Goden says “I can’t work with you until I’ve worked with you”

Mike Greenfield

  • Intro
    • @mike_greenfield
  • What is growth hacking? It’s simple
    • Choose the right goals
    • Find the tactics to meet those goals
    • Execute on them
  • Some growth skills
    • Product design
    • Copy Writing
    • Product Marketing
    • Coding
  • A data scientists role
    • does targeting
    • great targeting often beats everything else
  • Step one is usually (simple) rules-based targeting
  • Rules based targeting – Linkedin “people you may know”
  • Good input data are key
    • have you worked on a team together?
    • external data sources like address books
    • common connections
    • early on just use simple rules, maybe even a/b test the rules against each other
  • Next : calculate the probability that a user will perform a particular action
  • There are lots of ways to grow: data skills can help with a lot of them
  • Q & A
    • What models do you use?
      • Random forests – clusters of decision trees
      • He uses the python implementation
      • With circle of moms they automated the predictive model
    • What software do you use?
      • The strongest data scientists are good developers
      • They’re usually using R, SQL, python

Gustaf Alstromer

  • Intro
    • @gustaf
    • Voxer
    • AirBNB mobile growth
  • User Growth on Mobile
  • We’re creating experiences
    • Growth often becomes numbers – metrics conversions
    • But we’re interacting with real people and trying to create real value
    • We’re trying to create memorable experiences – the product we provide is to connect people with experience
    • I work on growth to make these experiences happen more
  • Amplify your magic moment
    • What can you do to help people experience a good moment with your product?
    • What can you do let that moment be shared by someone else
  • You can’t find mobile referrals like you can see referrers on the web
    • Traffic sources on mobile: advertising / search / email / sms / social -> app store -> ? where did they come from?
  • Voxer was a new behavior? What’s the existing behavior that matches? Walkie talkie
    • Let’s just try different names and see what banners people click on. Walkie talkie won by a huge margin
  • User funnel
    • Growth – sign-up
    • Activation – send a message
    • Engagement – send more messages
    • Retention – come back
  • Find “100 users who love you” – Paul Bucheit
    • They had thousands of users, but didn’t know who loved them
    • So they searched for users who were using it every day
    • So they had to ask “why do they love you?”
    • You can say “the product is good” but social products aren’t generally good in isolation, they exist in context
    • With Voxer, you have to have friends on the platform
    • What’s the difference between the active users and the people who don’t use it? The number of friends you have on the platform affects how much you’ll use the product
    • Internal dashboard: “Do we have 1,000,000 users? No.” (until it said “yes”)
    • You have to have a way for people to 1. be invited 2. find their friends 3. send a message
  • Inviting Friends
    • Measure
    • Unique users / % who tap next / % who tap “send”
    • a/b difference messages – shorter is better. Short personal and simple – almost something you would send to a friend. Then track it
    • % invite CTR / SMS -> Email conversion tracking
    • Are we viral yet?
    • Break the viral loop into steps, track the percentage who convert at each step, then you know where to optimize
    • Key Learning: Measure every single step in the funnel
  • Activation – Finding Friends
    • Match phone # with address books
    • Push notifications when friends join
    • Auto-create chat
    • = Amazing first time user experience
    • Having address books reduces your dependency on Facebook – not everyone is going to auth your app with Facebook
  • Social Graph matching:w
    • FB ids
    • phone numbers
    • email
    • match w/ graph -> fb friend list / mobile address-books
    • Your goal isn’t to collect people’s phone numbers, but rather to help them connect with their friends faster
  • Voxer’s growth
    • Your growth rate matters more than the absolute numbers
    • Soulja Boy – would sign up to Voxer, get 5000 notifications and quit the app (it’s not a broadcast app)
    • 13th most downloaded app
    • BUT it wasn’t all hacking
    • They started with a product that people loved using and optimized it after that
    • There was a big hole after nextel went away
  • Mobile Web Banners
    • “Smart banners”
    • Searching on google or getting an email lands you on the AirBNB site
    • What percentage of users that view the web install the app and open it? (Conversion)
    • The mobile web banners really helped the downloads take off
    • (Downloads are a vanity metric, but an okay proxy for value)
  • Tools
    • Yozio / Look back / Optimizely
  • Q & A
    • How did you get users in the early day?
      • Tried lots of things that didn’t work. Main thing that worked was focusing on active users and seeing what they did

Sean Ellis

  • Conversion Rate Optimization – Stacking the odds for growth
  • Intro
    • Invented the term “Growth Hacker”
    • @SeanEllss
    • Founder/CEO Qualaroo
    • First marketer at Dropbox / Logmein
  • The most important growth driver turned out to be word of mouth – That’s disconcerting for growth hacking
  • But there’s lots of factors that have to happen to drive engagement and deliver a great experience for people
  • Optimizing conversions is the most profitable growth driver in any company he’s been a part of
  • The Power of CRO
    • Optimization amplifies all customer acquisition channels
    • This makes all marketing more profitable
    • Doubling is not unusual in CRO
    • It opens up new profitable channels (because your conversion helps your ROI) and channels that weren’t viable before become viable
    • At logmein – after rigorous CRO they were able to make AdWords profitable
    • It improves customer happiness – people are able to do what they want to do more quickly
    • There’s always room to improve it:w
  • Pre-Req: Verify Product/Market Fit
    • You need to have a product people actually want in the first place
    • Until you have that, just have manual on-boarding
    • Survey.io – has some survey questions to help you find product market fit
    • CRO is about tuning a value delivery machine
  • Steps to better conversions:
  • Understand Visitors
    • People must consider your product a must-have. You can’t create a must-have experience if you don’t know what that experience even is
      • Visitors have intent and motivations
      • You need to understand blockages
      • E.g. how would you feel without the product? I’m looking to learn everything I can about the people that say they’d be very disappointed without our product
      • Dig in and find out everything you can from that person, try to find the kernel of benefit. What is the context? Why is that benefit important to people?
    • Intent
      • Then you can connect that benefit to intent
      • Intent is one of the most valuable things a marketer has to work with.
      • SMBs have one person cold-calling and walking in each day, it’s hard to break through.
      • We have a firewall that protects us from overload – intent is the way to break through that firewall
    • Example: Quizzle
      • Target 1st time visitors with a survey
      • Why are you trying to get your credit score?
    • Uncover points of friction
      • User testing (live or online)
        • User testing isn’t a great way to find out if your product is valuable
        • But if you give them a task and say “do XYZ” – you get to see questions and concerns and confusion. You can learn a lot about that experience.
      • Surveys to uncover conversion issues
        • “What’s the one thing that nearly stopped you from purchasing”
      • Example: $60k lost sales
        • What, if anything is stopping you from completing this order?
        • One particular browser was having a technical issue
      • Feedback sample size?
        • Just look for feedback that gives you inspiration about where to look
    • Decide on your CRO toolkit
      • funnel – Kissmetrics / mixpanel
      • a/b testing – optimizely / unbounce
      • user feedback – qualaroo / survey monkey
      • other – crazyegg / usertesting.com
      • Note: creating the test is often as hard as tracking the test
  • Prioritize Plans
    • List key customer intent (from research)
    • List friction to solve (from research)
    • Determine baseline conversions
    • Consider potential lift from tests
    • Script your first ten tests
    • Goals:
      • Max % that reach Must-Have experience
      • Desire – Friction = Conversion rate
      • Build desire, reduce friction
      • Hook Promise -> Must-Have experience
    • see: conversion-rate-experts.com
    • Apply to every on-ramp
      • Example: Dropbox
        • Homepage – sync
        • Folder share – collaborate
        • File share – send
  • Test & Analyze
    • Repeat the process frequently
    • [http://qualaroo.com/beginners-guide-to-cro]
  • Q & A
    • Anecdote: A longer credit card form converts better than a shorter form (for certain audiences) because it’s a higher degree of trust to go through the effort

Lara Klein

  • Introduction
    • @lauraklein
    • 3 Reasons they’re not converting
    • UX for learn start-ups
  • A conversion is taking the user to the next level, whatever that level is
  • Three reasons people don’t convert
    • They don’t understand what you’re offering
    • They don’t have the problem you’re solving
    • They don’t think you’re worth what you’re asking for
  • e.g.
    • That don’t get it
    • They don’t need it
    • They don’t want it
  • Five second tests
    • guerrilla usability
    • site interecept
    • Five seconds is plenty of time to figure out what something does
    • The reason they don’t get it is because you are not being clear
  • Test
    • Contextual Inquiry
    • Observational Testing
    • Customer Development
    • Learn about your customers lives, your customer’s jobs, your customer’s needs
    • Knowing about your customers is what allows you to solve their problems
    • You may not ask “does this solve a problem for you?”. People will say yes and they’re lying because people are polite.
  • Cost
    • Cost == Money + Time + Effort + Pain
    • Users don’t pay with just money
    • How can you tell if your product costs too much?
    • Pre-Build commitments
    • Usability tests
    • Price optimization
  • You can’t optimize your conversion if you don’t know which of these things you’re trying to fix
    • Changing your pricing page isn’t going to do anything if you don’t solve someones problem
  • Contact
    • laura@usersknow.com
    • @lauraklein
  • Q & A
    • How should you price your product
      • Ask users, generate comp analysis
    • If you have a product that requires education, how can you adjust the five second test
      • Does the product require education or do the benefits of the product require education?
      • People too often try to explain what the product does, not what the product does for me

Taking users from free to paid by Ivan Kirigin – YesGraph

  • Introduction
    • @ikirigin

    • Robots

    • YC08 Tipjoy

    • Engineer at Facebook Credits
      • used the same metrics and tactics on the FB credits team as the user growth (as well as ads)
    • Dropbox growth

    *

  • At FB: used the same metrics and tactics on the FB credits team as the user growth (as well as ads)
    • interesting that these procedures are the same across products
  • Freemium
    • If you have a paid product there is a barrier. The idea is that people can give you money later on
    • The biggest mistake is where you put what features
    • Often you’ll put a powerful feature in the pro side because its valuable, but you have a problem because you’re withholding a power of the service
    • Dropbox has the dimension of space
    • It’s not an obvious funnel from free -> paid -> enterprise
    • Enterprise is more about controlling and tools to use within a company
    • The freemium model has dividends because people in the company are already using it
  • Referrals
    • Dropbox grew a lot from referrals
    • BUT Dropbox has an amazing product. It can’t be emphasized enough – that took an incredible amount of work
    • It’s a lot of hard work. Find out the rough edges and smooth them
    • If you signed up from a referral, both of you would get space
    • You’re dealing with space, which is a virtual currency
    • Market norms vs. social norms – in ecommerce you’re not going to send a coupon to a friend to save money. You’re not trying to sell out your friendship
    • Symmetry – symmetric bonus – what’s more important is the invite. It lowers the threshold to send it because it doesn’t seem as spammy – you’re giving a gift. People aren’t yet looking for a bonus which they don’t use
    • The numbers of 2G / 250M / 50G – that works – you don’t want to cannibalize your subscriptions. You don’t want people to reach the paid plan level from referrals
    • Life-cycle – someone will use Dropbox, run out of space. Then they tell them to get more referrals. This happens several times – every time you run out of space. Its the whole life-cycle of the user – you want virality to happen throughout the whole life of the customer
  • Messaging
    • Dropbox messages users surprisingly infrequently – detail oriented quality bias
    • No newsletter. No new app announcements
    • No weekly newsletter about shared folders from other people
    • They’re really conservative and don’t want to bug their users
    • Messaging is about the excuses to reach out to the customer
    • /gs – a tour new Dropbox users go on
      • helps expose different features e.g. syncing
    • P.S. the cost is generally on the engineering side
      • the single most effective thing to referrals was adding P.S. to transactional emails
      • it was trivial to add, but huge lift to referrals
    • /edu – for students
      • students are cheap and have no shame so they spam everyone they know. They’re sharing homework.
      • they notified everyone with a .edu address about these programs
    • Spacerace
  • Enterprise
    • House ad – a huge chunk of growth around teams
    • Surprising and shows how much the freemium mattered – because it helps drive enterprise sales
    • Students after they graduate they get jobs and keep using Dropbox
  • /photos
    • one thing that was considered growth hack
    • automation – if you plug in your phone Dropbox will suck up the photos or videos
    • you just want it to be there
    • drives a lot of the right kind of engagement
  • Samsung & /gs
    • every photo and video is automatically in your Dropbox
    • people got the bonus space after they want through /gs and were on their way to becoming good users
  • YesGraph
    • Problem trying to hire
    • “Hey, who do you know who is good” – Usually it’s one or two people
    • Helps you scale referral recruiting
  • Flow
    • Signup / Job / Invite / Signup / Refer
    • An unspoken social network in companies – making it easy to invite the right people
  • Virality in the enterprise: It’s incredibly rare to have an enterprise product that has a viral growth factor
  • Q & A
    • Dropbox failures?
      • Many a/b tests didn’t work

Finding Paid Channels with Stan Chudnovsky

  • New User Channels
  • Viral / paid / Wom
  • Is paid for everyone?
    • Not really – your business has to have LTV
    • Good for: gaming companies, marketplaces
    • Not good for: media companies
  • When talking LTV every penny matters
    • The difference between sucks and great can be two cents
    • You must make a profit
    • Anything anybody can do to increase LTV is extremely valuable
    • If your LTV is positive relative to user acquisition you can spend more money to drive more users
    • When you are thinking about driving traffic and paid traffic, the good news is that every additional channel will add to the lifetime value
    • Organic + Paid = Power
    • LTV from viral consumers + the LTV from consumers you paid for == you can afford to buy
    • That is, if you buy a customer but they cause a viral customer then that increases what you can afford to buy
    • Zynga exploded because they did two things incredibly well
      • They cracked virality as well as they did
      • They supplemented with aggressive media buying (on FB in particular)
    • If viral factor = 0.2 LTV = $1 then you can pay $1.15 for cost of acquisition to break even
    • LTV:
      • Do you know it?
      • Can you measure it by channel?
      • Can you measure it by campaign?
      • Is Viral Factor from these users built into the LTV calculation
      • Per channel / campaign?
      • Does your FUE change based on channel/campaign?
    • People who are good at cracking virality are generally quantitative
    • Your target
      • Imagine a graph where the x axis is “meh phrase” “how phrase” and the y axis is “hot channel” “dead channel” – you want to be in the top right
      • E.g. AM radio might be a dead channel (but maybe not if you’re selling gold coins)
      • Discover the phrase
        • The right phrase can make the difference between buying users in a profitable way
        • “unlimited DVD rentals”
        • “lose 20 pounds in 20 days”
        • “free IQ Test”
        • “play dragons of Atlantis”
        • The companies that figure this out are doing exponentially better than the ones that don’t
        • Find the phrase that people are psychologically given to, for whatever reason
    • Paid channels come and go
      • You have to find the phrase that works on that channel
      • AOL / Yahoo / MSN
      • Remnant Web Banner Networks
      • Mobile Remnant Ad Networks
      • Pop-Ups / Pop-Unders – perform incredibly well in the beginning, but then they just stop working – 10 years ago that was how it worked. Conversion was like 5x of the competitors
      • Google SEM – it worked because it was cheap, but it becomes more and more efficient so it’s harder to buy
      • You have to be always ready because new opportunities present themselves and you have to be ready to hop on them
      • Google AdSense
      • Myspace – fast an effective
      • Facebook Web
      • Facebook Mobile – 4x price after 6 months – was cheap for less than 6 months
    • Instrument 3 buckets
      • Short term ROI 3 – 6 months
      • Longer term ROI 6 – 12 months
      • Longest term ROI 12+ month
      • TVs can be instrumented too
    • Q & A
      • How do you decide between in-house or using an agency?
        • I haven’t found an agency that knows how to do it
        • They’re not setup to find how to make it work for the company
        • They say “what budget do you have setup” and they spend it
      • What percentage of acquisition costs compared to the LTV is the right range?
        • Depends on the margin of the business
        • Between 3% or 20%
      • How do you get started?
        • At first, just use LTV of a shorter time period (3/6/9 months) – after you’ve been running for 12 months you can use more aggressive numbers
      • How do you keep track of new channels that are coming out?
        • Genuinely no one will tell you. That’s one of the benefits of being in Silicon Valley
      • What do you see as the next big channel?
        • Twitter is working for some people
        • G+ identity

How to drive viral user acquisition from zero to 1million Simon

  • Introduction
    • SpeedDate.com
      • 20M users
      • 100k subscribers
    • Network of Facebook apps of 150mm users
    • Ipsy
  • Create an awesome product first
    • The first step to 1 million users is to build a product that people care about
    • jdate – I’d hardly hear back
    • Once in a while I’d meet someone. But they were totally different than their profile
    • Hard on-boarding
      • Static profile listings
      • Compatibility matching
      • Real-time matching
        • Seconds within signing up you’re connected to real people
    • 5 things
      • Wow/trigger point and other metrics
        • What’s the best experience that someone can have on speedate while they’re online?
        • Not subjective – let’s really quantify it.
        • Real-time connection -> reply metric -> overall experience
        • Measuring if someone actually had a conversation
        • These were called the WOW metrics
        • How can we increase the percentage of users that get to the WOW experience?
      • User needs
        • Every user has expectations
        • Expectations: Clear UX / Real people / baseline relevance (around location, around age)
        • Desires: Control / Context / Preference relevance
        • Unmet needs: what do other channels lack that we can offer?
      • Channel profitability
        • LTV – Cost/sub – Payback
      • User feedback
        • uservoice / foresee / customer loyalty feedback / usability testing
      • Competitive research
        • eHarmony / match.com
    • Metrics for awesome products
      • Addicted customers
        • high percentage return daily / weekly
        • 30% dau mau
      • Customer satisfaction
        • 4+ star rating
        • 40% if product discontinues
        • would send a survey to 1% of users everyday and track this metric
      • Significant word of mouth
        • 50% organic traffic
      • Breakout economics
        • 250% scalable ROI
  • IPSY
    • Subscription sampling + impulse buy + community
    • Give people tools to share more
    • People are excited to get this bag of makeup so they’re excited to share it.
    • So first: you need an awesome product before you try to move to the growth phase. (And make sure you’re measuring it)
  • What are the viral channels?
    • FB / Twitter / Instagram / Linkedin / SMS / Email
    • These are best known, but they’re not the only ones
    • Your product might use many of these but start out with one
    • Product in their early days allow for growth that more mature platforms don’t
    • The first step in getting viral growth is picking the platform you’re going to use
  • Viral growth
    • The viral growth itself isn’t as important as the metrics that make it up
    • invite * invites per inviter * signup from invites
    • 20% invite x 200 invites per inviter x 2.5 signups = 1
    • If you have more retention than engagement you don’t have to do it in the signup flow, but people typically put the invites in the signup flow
    • After you have invites sent out, you have to make sure people actually signup and invite again. This is called “closing the viral loop”
  • Viral user acquisition analytics
    • a/b/ testing system for every conversion
    • setup the analytics for per-channel analysis of viral metrics
  • Lean Development Process
    • Ideation -> MVP -> QA -> Daily deployment -> AB test -> Measure and Iterate -> Ideation again
    • AB test all features
    • You don’t know what’s going to work. AB tests help you know objectively what’s going to work
    • (And usually only 1 in 10 things work)
  • Summary
    • create an awesome product first
      • address a big user need that isn’t being solved
      • measure to know if it’s really working
    • Only once you have that can you drive viral growth
  • Q & A
    • FB platform in 2008 – they were launched in 2007
    • I think it would be interesting to see what successful companies were launched by growth from FB in that time

Supply Hacking – lessons from building Udemy and Lyft

  • Intro
    • @gaganbiyani
    • Working on Sprig
  • Tradecraft
    • We started this conference because most conferences don’t speak to the specific tactics used to grow companies
    • Tradecraft is starting DevBootCamp for people who want to become great at sales, UX, and growth
  • The problem:
    • When you’re trying to get a million users to use your app, these things we’ve been talking about apply
    • But its an entirely different challenge to get supply side on board when you’re building a marketplace
    • Udemy instructor don’t know each other and only 1 in 1000 are even qualified
    • For b2b companies the problem is the same.
    • So how do you build a process to win in those situations?
  • Sprig – healthy meals on-demand for $12
  • If you have a chicken and egg problem, fake the chicken
    • Udemy: take the open-source youtube content and put it on the site
    • Lyft: you can create or pay for the supply side of your company
    • Your product probably isn’t perfect in the early days, so working directly with your supply side helps you see how your product works
  • Or make your own chicken
    • Lyft: they paid drivers to sit on the road idle
    • Udemy: made their own classes
  • For users, effort and dollars are related
    • It’s far more valuable to go in and create the exact perfect experience for the user (at a loss) then try to create the experience from both sides at the same time
  • Early success stories
    • The whole company would just share the couple of courses that were good
    • If you’re doing b2b, your whole company job is just to make two customers successful and get case studies
    • Then promote the success of the courses / customers / drivers
    • Money talks – pitched “make $28/hour”
    • For Lyft, made testimonials and interviews and videos of the drivers and then put that on the site
    • The first Udemy copy was “create an online course”, then it was “create a course and make money” then “tell everyone courses make money”.
    • It’s important to know how important the order of operations is at a startup
  • Scale is all about leverage
    • ToutApp – email management
    • Pipedrive
    • outsource labor to email people who they think would be good at making courses
    • One set of assistance that would google “learn python” “learn guitar” and create a massive email list of teachers of that subject
    • Another group of people would customize that email and send a customized individualized email “Hey have you ever thought about creating a course on Udemy”
    • This works better because a person feels they’re getting a personal email
    • This is the virtual assistant strategy – hire someone to go do a large amount of the work that you’re doing yourself
    • First you have to do everything yourself and then slowly slice it off to someone who is cheaper.
    • Onboarding at Lyft is now all video. (It was live previously)
    • At Udemy we would upload it for them
  • Constant supply/demand balance
    • If you’re lucky, you have to keep balancing it
  • Contact
    • @gaganbiyani
    • gaganbiyani@gmail.com
  • Q & A
    • See: Pipedrive & ToutApp
    • Thumbtack and ToutApp
    • How do you validate the supply that you’re getting?
      • Quality control on supply side
      • if someone has a bad experience you’re not coming back
      • in you then have to compliment that funnel with a strict quality control mechanism

A Personal Finance Approach to Growth – Andy Johns

  • Intro
    • Director of Growth at Wealthfront
    • Quora
    • Twitter
    • Facebook
  • Trying to give a framework or a mental model
  • My favorite mental model
  • Core finance principles
    • Increase your income, decrease your debt
    • Start investing early
    • Diversification is mandatory
    • Invest in cost efficient assets
    • Compound interest is your best friend
  • Building growth in users is no different than growing your retirement account
  • If you imagine two lines on a graph, income goes up, debt goes down. Whatever amount is left over can be invested
  • Invest in things that are cheap
  • Startups
    • Increase growth, reduce product / tech debt
      • A lot of startups hire a lot of people without applying them with focus to the things that matter
      • e.g. we don’t apply them on the things that have caused them to grow
      • you join a hot growing startup 20–200 and build a bunch of new features and make it almost unusable
      • so when you’re on the cusp of hockysticking, you need to invest in growing what you have today
      • move features do not mean you’ll grow more
      • e.g. you didn’t need photo filters to grow twitter
    • Start growth projects early
    • Diversify your growth channels
    • Invest in cost efficient growth channels
    • Focus on rate of growth, not absolute growth
  • FB vs. Myspace
    • Growth rallied around the idea that everything they do will be around that metrics – 10 friends in 7 days
    • Myspace did the opposite -> you can customize your profile. Made it atrocious
    • Fast, simple, predictable
  • Start growth early
    • compound interest
    • you need to prove some value within 18 months, 36 months
    • your ability to prove your growth influences if you’re going to be able to raise money
  • Diversified growth channels
    • Efficient frontier
    • Graph – X: Risk low -> high. Y: return Low high
    • Same Risk more return
    • You are taking a risk by doing a startup so you need to take control of your growth
  • Invest in cost efficient channels
    • There is a positive correlation between the successful startups today and their lack of need to pay Facebook
    • Easy growth doesn’t mean the best growth
    • If you’re building a two sided marketplace, get ready to build it by hand for years
    • Uber built a playbook, a road map for SF. So they’ll hire a GM in each city and have them replicate that
  • Focus on your rate of growth
    • 2% weekly growth – everything we did we could compute the growth
    • When you sit down at a company and wonder “What am i going to do?” that’s what you’re going to do – (grow those numbers)
    • The only thing I’ve done is doggedly focus on one metrics longer than other people are willing to do so
    • P/L – profit and loss
    • This is what you do for a living – optimize this metric. Have a ridiculous focus on a metric that has meaning for your building
    • 5% weekly growth rate
  • Growth hacking is misleading
    • If you can find a 10% improvement every 4 weeks, then you can grow
    • That is your job, set yourself up to find out what those things are
    • You cannot sustainably grow something that sucks
    • Find something that works, then pour fuel on the fire, that is your job
  • Anecdote from twitter
    • A more scientific approach to product met a product religion
    • Decided to do what he thought was right for the business even if he could get in trouble
    • He and a designer and engineer they hacked together a solution and rolled it out to 10% of all traffic for 2 days
    • Data was overwhelmingly positive
    • Got 60–70,000 more signups per day – biggest growth increase in the companies history
    • Sometimes it takes that willingness to decide that this is the right thing to do
  • Product first
    • color
    • Myspace
    • cuil
    • viddy
      • 40mm active users
      • b/c facebook opengraph integration partner
      • financing didn’t check how well their users were engaged
      • growing via a single channel that you can’t control or predict is the same as putting all your money on one stock
      • I work on products that can grow on a dozen different channels
      • Your values is proportional to the growth channels which you directly control
      • Mobile has fewer channels, it’s harder to measure
  • Some books
    • A random walk down wall street
    • the little book of common sense investing
    • Apply finance ideas to growing startups
  • Build a growth table
    • if you don’t know what source contributes what amount to your traffic then you’re throwing darts blind
    • Source || % of traffic || CVR || Absolute % of Growth
    • Know your proportion of your traffic and what it contributes to your growth
    • Compare to finance – they track the capitol in and out of your business
      • Consumer startups have to have a bunch of users. So you grow and then on the heels of that you make money
      • Finance looks at the flow of capital but we also don’t have a growth team tracking the flow of users
      • Finance team is responsible to report expenses by channel
  • Summary
    • Growth is important above how sexy or cool it is right now
    • They’ve becoming better and better at a fundamentally valuable thing they can do
    • In order to not repeat what happened in the 90s we have to be better at building businesses
  • Q & A
    • What if you’re small?
      • In the model of growing weekly you have variance, seasonality, or the numbers are small. You can smooth it out by monthly, but in any case you can track the impact of new features on growth rate
      • Education space products have this – in the summer time there is nothing going on here because school is out
      • Just apply averages: what is my average weekly growth rate on a rolling 4 week period
    • How do you keep people motivated on one metric?
      • Results lead to excitement.
      • The sooner you can find your way to things that make results, the more excited people get. That buys you excitement into the next week where you have to grind it out again and do that once more.
      • If you don’t have the team or tools and it takes 6 months before you can produce your first result then that’s tough
      • One of the first things at Quora is he got some trophies from north Dakota and it had a battleship and gave it to the team
      • 100MM active users -> 300MM in 12 months (no ones done that before). If they hit it, they would pay for a 3 day trip to Vegas
    • Importance of multichannel attribution
    • How do you weight priorities of things that produce growth but they require a lot of engineering effort?
      • You need to have a bias towards action. Idea / description / baseline stats / estimate
      • What will the new growth be if I do it
      • LOE 1 hr / 1 day / 1 month
      • Keep it simple and estimate the growth impact
    • Mobile attribution space
      • Yozio
      • Hazoffer – mobile app tracker
      • Examples: Messaging app – a meaningful percentage of growth came from invites. Track invitations that came from each invite point. Tracked the invitation point with the invite
        • Can do the same thing for paid campaigns
        • Yozio click tracking on anything
        • The remainder is just raw organic app store rankings
    • Failures
      • We’re basically trying to find how wrong we are as fast as possible
      • Just have a good system for improving your pattern recognition in the future
      • 70 experiments in 60 days and tripled the growth rate. Maybe only 20–30% of experiments worked at first.
    • How does a growth team fit in a company
      • Enterprise has a senior staff sales guy
      • Growth is a horizontal layer above/below product
      • Growth can come from any part of the product. But it’s not necessarily the product commander’s role to think about the growth affects of a particular part of the product.
      • The growth team should be able to touch any part of the product if they want to
      • Growth engagement and mobile adoption
      • The team moves throughout the company and helps improve adoption for different areas
Posted in programming | Leave a comment

Pebble Watchface: Fireflies

I recently received my Pebble watch and I’ve created a couple of watchfaces for it. The one I want to share with you today is a study in swarming behavior: Fireflies

ScreenShot

Click to see animation

ScreenShot

Download v1.0 here

You can join the conversation in the Pebble Watchapp Directory and, if you’re interested, I’ve released the source on github

Posted in programming | 1 Comment