daft punk
halo warrior
jedi knight
search this blog
Sunday, November 2, 2008
Saturday, May 17, 2008
Friday, March 28, 2008
engagement?
what the hell does 'engagement' mean?
http://www.kaushik.net/avinash/2007/10/engagement-is-not-a-metric-its-an-excuse.html
1. why does website exist?
2. define 3 or less metrics to measure success (accomplish purpose)
3. revenue/conversion; engagement - how to measure?
4. don't call that metric 'engagement'
ideas....
1. are you engaged with us?
2. would you recommend site to friends?
3. primary market research
4. customer retention over time
5. recency of unique visitors (last resort)
6. fraction of multi-visit visitors (> 1 visit)
do visitors search the site?
do they 'save' stuff?
do they sign up for email?
how many entries to home page?
http://www.kaushik.net/avinash/2007/10/engagement-is-not-a-metric-its-an-excuse.html
1. why does website exist?
2. define 3 or less metrics to measure success (accomplish purpose)
3. revenue/conversion; engagement - how to measure?
4. don't call that metric 'engagement'
ideas....
1. are you engaged with us?
2. would you recommend site to friends?
3. primary market research
4. customer retention over time
5. recency of unique visitors (last resort)
6. fraction of multi-visit visitors (> 1 visit)
do visitors search the site?
do they 'save' stuff?
do they sign up for email?
how many entries to home page?
Sunday, March 23, 2008
easter sunday march 23
roy williams is a great coach, first at kansas, now at north carolina.
memphis and ucla survived. north carolina is whipping arkansas. kansas had a close game but pulled away late. north carolina has looked most impressive as the number 1 seed.
stephan curry, son of ex-nba player dell curry, put on an impressive performance against gonzaga and georgetown. davidson may be for real... we'll see next weekend.
sweet 16 is set
memphis and ucla survived. north carolina is whipping arkansas. kansas had a close game but pulled away late. north carolina has looked most impressive as the number 1 seed.
stephan curry, son of ex-nba player dell curry, put on an impressive performance against gonzaga and georgetown. davidson may be for real... we'll see next weekend.
sweet 16 is set
Monday, March 17, 2008
random rants
1 american dollar is worth 99 canadian cents... what the fuck, what is up with that?
the houston rockets have won 22 in a row... 2nd best in history (lakers, 33 in 1972)... i like tracy mcgrady and yao ming (injured)... keep it up... this is truly an amazing streak when you consider that bird, magic and jordan never won this many in a row
the houston rockets have won 22 in a row... 2nd best in history (lakers, 33 in 1972)... i like tracy mcgrady and yao ming (injured)... keep it up... this is truly an amazing streak when you consider that bird, magic and jordan never won this many in a row
Thursday, February 28, 2008
Thursday, February 21, 2008
google website optimizer
--->tom leung gwo ama vki
---->vki studios feb 5 ama testing
stop guessing, start testing, increase conversion!
Labels:
a/b,
gwo,
multivariate,
mvt,
test,
testing,
web analytics
Wednesday, February 20, 2008
phases of a business
phase 1: startup, make some money
phase 2: cut expenses
phase 3: focus on building value
phase 4: cash out (sell)
what they still don't teach you at harvard business school, page 90
phase 2: cut expenses
phase 3: focus on building value
phase 4: cash out (sell)
what they still don't teach you at harvard business school, page 90
moving average
look at 30, 60, 90 and 360 days
daily: current day, previous day, 30 day average, 90 and 360.
daily: current day, previous day, 30 day average, 90 and 360.
check out crazy egg
check out crazy egg.
http://www.youtube.com/watch?v=Qr-PVnPmqVo
heat maps offer more than site overlay.
http://www.youtube.com/watch?v=Qr-PVnPmqVo
heat maps offer more than site overlay.
Monday, February 18, 2008
2 metrics plus 1 (jim novo)
http://www.google.com/analytics/cu/ac_two_metrics_for_you.html
1. bounce rate
2. bounce rates for top entry or landing pages
3. percent 1 minute visits
track these percentages over time (i.e. excel graph)...
1. bounce rate
2. bounce rates for top entry or landing pages
3. percent 1 minute visits
track these percentages over time (i.e. excel graph)...
Sunday, February 17, 2008
measure effectiveness of your web pages
excellent analytics tip #11 by avinash kaushik
5 step program:
1. don't obsess about your home page
graph home page visits over time (y axis is % visitors)
what % of site traffic actually views your home page?
2. compute your cliff, then jump
this applies if a 'small' number of pages accounts for 'most' of your traffic. for ticketing sites such as ticketmaster and razorgator, i have not found this to be the case.
3. bounce rate - it's good for you
what is wrong with the page?
content, calls to action, navigation?
who is coming to the page?
referring urls, campaign id's, search keywords, etc?
expired promotion?
20% (ok), 35% (cause for concern), 50% (worry)
4. site overlay - something to love
why is page underperforming?
where are people clicking?
5. think holistically - multiple metrics, key context
you actually have to interact with your web pages to pinpoint problems.
bonus: insight -> improvements: multivariate testing rocks
test all ideas, don't go live with just one.
test different images, calls to action, content and layout that works best for your customers
tools: offermatica, optimost, sitespect, google's website optimizer
javascript tags on the page
5 step program:
1. don't obsess about your home page
graph home page visits over time (y axis is % visitors)
what % of site traffic actually views your home page?
2. compute your cliff, then jump
this applies if a 'small' number of pages accounts for 'most' of your traffic. for ticketing sites such as ticketmaster and razorgator, i have not found this to be the case.
3. bounce rate - it's good for you
what is wrong with the page?
content, calls to action, navigation?
who is coming to the page?
referring urls, campaign id's, search keywords, etc?
expired promotion?
20% (ok), 35% (cause for concern), 50% (worry)
4. site overlay - something to love
why is page underperforming?
where are people clicking?
5. think holistically - multiple metrics, key context
you actually have to interact with your web pages to pinpoint problems.
bonus: insight -> improvements: multivariate testing rocks
test all ideas, don't go live with just one.
test different images, calls to action, content and layout that works best for your customers
tools: offermatica, optimost, sitespect, google's website optimizer
javascript tags on the page
Monday, February 11, 2008
getting started web analytics - first step - glean macro insights
getting-started-with-web-analytics-step-one-glean-macro-insights
1. how many visitors are coming to your website?
2. where are visitors coming from?
3. what do you want visitors to do on the website?
a. why does website exist?
b. what are your top 3 web strategies that you are working on?
i.e., paid, affiliates, update content, effective merchandising, etc.
c. what do you think should be happening on your website?
4. what are visitors actually doing?
a. top entry pages
b. top viewed pages
c. site overlay (click density)
notice we have not mentioned conversion rate and path analysis
1. how many visitors are coming to your website?
2. where are visitors coming from?
3. what do you want visitors to do on the website?
a. why does website exist?
b. what are your top 3 web strategies that you are working on?
i.e., paid, affiliates, update content, effective merchandising, etc.
c. what do you think should be happening on your website?
4. what are visitors actually doing?
a. top entry pages
b. top viewed pages
c. site overlay (click density)
notice we have not mentioned conversion rate and path analysis
Saturday, February 9, 2008
4 attributes of a great metric according to avinash
1. simple
2. relevant
3. timely
4. instantly useful
example of a great metric: bounce rate
simple, tells you where you are wasting marketing dollars, which pages suck
3 important final lessons...
1. perfection is the enemy of good enough
90% is good enough. follow best practices, then move on. an educated mistake is better than no action at all.
2. critical few
which 3 metrics define your existence?
3. metrics life cycle
define -> measure -> analyze -> action -> improve or eliminate
do this once a quarter
2. relevant
3. timely
4. instantly useful
example of a great metric: bounce rate
simple, tells you where you are wasting marketing dollars, which pages suck
3 important final lessons...
1. perfection is the enemy of good enough
90% is good enough. follow best practices, then move on. an educated mistake is better than no action at all.
2. critical few
which 3 metrics define your existence?
3. metrics life cycle
define -> measure -> analyze -> action -> improve or eliminate
do this once a quarter
web analytics demystified - winner of a semmy!!!!!!
congratulations avinash!
1. get primitive basics out of the way
look at how these trends change...
visits
pageviews
pages/visit
bounce rate
avg time on site
% new visits
2. understand traffic sources
direct traffic
referring sites
search engines
other such as email
3. fix stuff/save money
bounce rate for top entry pages
bounce rate for search key words (not as helpful in my opinion)
google website optimizer for a/b or multivariate tests
4. site overlay
click density
segment clicks if possible (can g.a. do this?)
5. focus on outcomes ("to get rich is glorious" - deng xiaoping)
what is purpose of your website?
e-commerce?
revenue, conversion rate, transactions, aov, products sold
are you converting enough?
what is selling and what is not?
why?
how much of it?
segment traffic: paid vs organic, campaign vs direct, new vs returning, geographic locations, etc
1. get primitive basics out of the way
look at how these trends change...
visits
pageviews
pages/visit
bounce rate
avg time on site
% new visits
2. understand traffic sources
direct traffic
referring sites
search engines
other such as email
3. fix stuff/save money
bounce rate for top entry pages
bounce rate for search key words (not as helpful in my opinion)
google website optimizer for a/b or multivariate tests
4. site overlay
click density
segment clicks if possible (can g.a. do this?)
5. focus on outcomes ("to get rich is glorious" - deng xiaoping)
what is purpose of your website?
e-commerce?
revenue, conversion rate, transactions, aov, products sold
are you converting enough?
what is selling and what is not?
why?
how much of it?
segment traffic: paid vs organic, campaign vs direct, new vs returning, geographic locations, etc
Friday, February 8, 2008
how to increase site conversion rate? (courtesy of avinash)
1. segment, segment, segment
segment by referral source, landing page, search terms, etc.
2. look at average order size, visitors, bounce rates in conjunction with conversion rate
3. put context around conversion rate
i.e. lowering price would most likely increase conversion rate but is that a great thing if you make less revenue?
4. track abandonment of each page (or every important page)
5. what is your true audience?
Testing...
method 1: a/b testing
simple, easy but limited
method 2: multivariate testing
more variables but more complicated
offermatica: marketing focused
optimost: technical
sitespect: no javascript code
google website optimizer: free with adwords
method 3: experience testing
Thursday, February 7, 2008
5 tips for increasing sales online (courtesy of avinash)
http://www.avangate.com/interviews/avinash-kaushik_7.htm
1. where is your traffic coming from?
a. referring urls
b. search engine keywords
c. campaigns
2. don't obsess over your home page
instead, look at your top 50 entry (or landing) pages. which entry pages have bounce rates > 50%?
3. don't obsess over conversion.
no matter how hard you try, overall site conversion will be around 2%. focus more on enhancing visitor experience.
4. why do people visit your site? (conduct survey, 3 fundamental questions)
a. what is the purpose of your visit today?
b. were you able to complete your task today?
c. if you were not able to complete your task today, why not?
5. use a robust multivariate testing tool such as google's website optimizer (free with adwords) or offermatica (expensive)
1. where is your traffic coming from?
a. referring urls
b. search engine keywords
c. campaigns
2. don't obsess over your home page
instead, look at your top 50 entry (or landing) pages. which entry pages have bounce rates > 50%?
3. don't obsess over conversion.
no matter how hard you try, overall site conversion will be around 2%. focus more on enhancing visitor experience.
4. why do people visit your site? (conduct survey, 3 fundamental questions)
a. what is the purpose of your visit today?
b. were you able to complete your task today?
c. if you were not able to complete your task today, why not?
5. use a robust multivariate testing tool such as google's website optimizer (free with adwords) or offermatica (expensive)
Friday, February 1, 2008
Subscribe to:
Comments (Atom)