SXSW: DAY 1

Since I skipped the entire first day, including all the parties, I was able to get started bright and early on Saturday morning.  First thing I had to do though was pick up my badge at 9 am.  There were quite the number of people who were there before 9 waiting to get their badges as well so I was a little worried that I would be spending a good amount of time just getting that done.  Thankfully, the process went much quicker than the security line at O’Hare airport went the day before.

Anyhoo, these are the sessions I attended:

Session 1: Using Twitter to Improve College Student Engagement
Dr. Rey Junco, @reyjunco

As one of the niche sites I work on is specially related to the college market, I thought this would be a valuable session to attend.  The content was a little different than I had expected since it was more geared to engaging students in an educational setting, not for general business.  I still found it interesting nevertheless.  The speaker was definitely quirky but in a good way.  He was kind of like one of those professors that borderline being really cool and really cheesy at trying to relate to his younger audience.

While the studies in this session were mostly concentrated around twitter, facebook was discussed briefly in the beginning.  One interesting point I thought that was brought up was that time spent on facebook is a negative predictor of how much one is engaged in facebook.  Positive predictors are actually events, comments and viewing photos.  In fact, less engaged students spend more time on facebook.  Apparently, getting metrics on facebook is difficult and faculty was more interested in integrating twitter over FB.

The study conducted 2 experiments:
Study #1: In the experimental group, students were actively told to engage with twitter by the faculty.  In the control group, students were actively told to engage in ning by the faculty.

Study #2: Students receive no active encouragement to use twitter
Rather than spend a huge chunk of my time writing up the activities and results of each of these studies, I found you can actually download some of the slides used in his presentation and even more detailed information about the experiments on his blog: http://blog.reyjunco.com/

Twitter Improves Student Engagement and Grades

The studies found that faculty must use twitter to:
1) Engage with students
2) Integrate course content
3) Encourage collaborative learning

Another takeaway I thought that was interesting about this session was a specific technology used during the study.  The software was called Leximancer and it was used to link concepts across the various tweets of the students and faculty used in the study.

Session 2: Designing For Silence : Using Email for Good
Michael Jackson Wilkinson, Posterous, @whafro

I thought this was an interesting session to attend since we send a lot of email to our users and you always wonder what’s the right amount of emails to send before it becomes detrimental to your business.  You have to remember that not only are you sending emails, but everyone’s inboxes are being filled with groupon, livingsocial, facebook, twitter and all sorts of sites you might be signed up for.  So how do you keep your company emails from being lost in all the clutter and so that your users don’t end up unsubscribing from your emails all together.

“Strive to make noise only when it improves the silence.”

Bac’n (pronounced: bake – on)

  • Spam you sort of want to get. (haha)
  • Examples: Amazon purchase confirmations, twitter direct message notifications, etc.
  • Insight: Not everything everyone follows (such as every friend on facebook) is equal.

Things you can do to modify current email interactive with users before you see diminishing returns:

Bundling

  • Allow users to elect to wrap multiple notifications into oneExample: Daily digests versus instant notification of each interaction

Bundling Wisely

  • What should you bundle?
  • One idea: Customize on the frequency for each person or group of people.
  • Generally users give you clues on how frequently they want to interact with service such as how often/when they login

Highlighting Favorites

  • Reorder email content based on frequency of views
  • How much time are they spending on certain items

Since You’ve Been Gone

  • Give users a nice recap of things that have happened since they last visited the site
  • Example: Most active things others have been viewing
  • Example: What were the last 5 things I viewed the last time I was at the site.  You can have the browser store these items.

Overall Lessons:

1. Don’t forget the important stuff

  • Don’t leave out the detail to get users to go back to the site.  Generally backfires and ends up annoying users more than helping your site.
  • Example: Facebook photo comment notifications don’t tell you which photo your friend commented on

2. Don’t beat around the bush

  • If there is an action required, make sure it’s clear to the user
  • Be clear with what you want users to do with transactional email messages
  • Example: Big button for link you need users to take action upon

3. Meta data

  • Keep app or company name in the from line
  • Keep the from line to 20 – 25 characters
  • Specific subject lines

4. Effective messages are designed messages

  • This doesn’t necessarily mean they need to be the best looking emails
  • It more means that you should consider the hierarchy of what you want the user to do

5. Be careful with promos

  • Keep it so there is not more promos than content or that promos do not confuse user on what to do

Session 3: Metrics-Driven Design
Joshua Porter, @bokardo

This was probably in my top 3 sessions of the entire conference.  The great thing about this session was that the presenter gave everyone a link at the very end where people could download his slides: http://bokardo.com/talks/metrics-driven-design (I wish I had known this otherwise I wouldn’t have been furiously trying to write down every word).

These are the key things I took away:

Optimization asks: What works best in the current model?

Design innovation asks: What is the best possible model?

Metrics are simply the numbers that measure the effectiveness of your business.

  • Example: “Must improve conversion by… “

5 Reasons Why Metrics are Designers’ friends:

  1. Metrics reduce arguments based on opinion
  2. Metrics give you answers about what really works
  3. Metrics show you where you’re strong as a designer (and weak)
  4. Metrics allow you to test anything you want
  5. Clients love metrics

With metrics, you are measuring how you are moving people along the usage life cycle.

There are 5 types of metrics:

  • Acquisition
  • Conversion
  • Engagement
  • Satisfaction
  • Emergent

Acquisition Metrics

  • CPA – Cost Per Acquisition
  • If your CPA is higher than your lifetime value, you’re in trouble
  • Acquisition vs. Referral
    • Example: DropBox
    • Paying too much for Google AdWords
    • Created 2 sided incentive program for referrals where both the referrer and referree benefit
    • Extremely successful
    • Increased signup permanently by 60%
  • Performable acquisition metrics include:
    • Comparative metrics
    • Revenue by channel
    • Revenue by keyword
  • Email lists then to be extremely valuable

Conversion Metrics

  • Example: Twitter redesigned their sign-up flow
  • Looked at their most active users and re-engineered what they were doing
  • Made their site very topic based
  • They added a 2nd page – a category list where people can self-select what to follow

Engagement Metrics

From most traditional down:

  • Hits
  • Page Views
  • Visits
  • Unique Visitors
  • Returning Users
  • Registered Users
  • Customers
  • Frequency
  • Time on site
  • Daily active users

Cohort Analysis

  • Break up users into segments based on when they started

Prevention: Facebook Deactivation

  • Added “These people are going to miss you” section
  • Change accounted for 1 million people NOT leaving

Satisfaction Metrics

  • How likely would you recommend our company to a friend of colleague? (Net Promoter Score)
  • Score = % Promoters – % Detractors
  • Mint.com did not have a high viral score but had a high NPS

Emergent Metrics

  • Not obvious at first by more over time
  • Example: FriendFeed
    • They discovered magic number is 5
    • Introduced a way to easily add friends
  • Example: Blogger
    • Realized most critical piece/metric was the number of posts

Principles of Metrics-Driven Design

  1. Optimize in small steps; innovate with daring leaps.
  2. No design survives contact with the user.
  3. Small improvements, taken together, yield amazing results.
  4. Testing is empowering, reversion is cleansing.
  5. Metrics are not creative: human beings are.
  6. All team members are responsible for the user experience.
  7. If metrics aren’t actionable, they aren’t useful.
  8. Design is never done.

champ

Leave a Reply

Your email address will not be published. Required fields are marked *