Conversion Optimization from the Business Perspective:
The Dollars and Cents Behind It All
!
I’m the business guy that understands technology.
Neon Rain Founded in 2002 Clients from Denver to Singapore
The piecesTraffic
Conversions
Retention
SEO / PPC
Offer
Unique Service
The piecesTraffic without Conversions and Retention
Open drain
The piecesTraffic with Conversions but no Retention
You have to keep the inflow greater than the outgo
The piecesTraffic with Conversions and Retention
What’s your focus? Stop looking at the wrong metrics
i.e. Trying to rank #1 for keywords with no traffic
The right metricsDefine what matters
Let’s start with a riddle3 women are having dinner and the bill comes to $30. Each of them puts in $10. 3 x $10 = $30
As the waiter rings them up, he realizes it should have totaled $25.
The waiter starts to take the $5 back to the women but realizes that he can’t split the $5 over 3 people so he pockets $2 as a tip and gives each woman back $1. Therefore, now they have paid $9 each.
$9 x 3 women = $27, now add in the $2 the waiter kept and that totals $29.
So, what happened to the missing dollar?
BaselineKnow your numbers: Goals / Unique Visitors = Conversion rate
Example: You spend $2,000 / month on PPC and get 10,000 impressions and 500 clicks. Your initial conversion rate from PPC to click is 5% and your cost is $4 / click.
On your website you can see that 1 in 10 contacts you. Now your conversion rate is 10% of the clicks or 50 contacts out of 500 clicks.
Even further down the line you know that you close 5 new clients from your sales process. 5 out of the 50 contacts is again, 10%.
If we look up the chain we can see that out of 500 clicks you closed 5 new clients. That is a 1% conversion rate. You also know that your acquisition cost is $400. (You spent $2000 and got 5 clients)
Even better, add in your average lifetime sales value to the equation.
The ProcessIf you can measure it, you can change it.
Then test and measurecontinually.
Example of a metric
This particularclient sells ink
Example of a metric
This client sellsreplacement inkcartridges.
Getting to the coreThey made money off their email marketing. They would make an average of $1 for each person on their list.
The goal: Get more people on the list
Tweak itTo increase purchases they would send “perfectly” timed email.
Just before their ink ran out.
Data was based on historical purchases.
Their “metric” was email sign ups
Ways of testingA/B Testing (Split test)Testing one component with 2 different options - including a control
Multivariate TestingTesting multiple components with many different options - including a control
Example of A/B TestText initially said “Buy Now”Baseline: ~3% bought
Changed to solid backgroundwith no significant changeChange: Not measureable
Changed to “Add To Cart”Change: ~+3%
Example of A/B Test
Increased % of visitors using wishlist by ~+43%
Example of A/B TestIncreased % of visitors using
wishlist by ~+43%
Oops % of people buying down by -1%. Net of +2% now
Example of A/B TestWe Decided to leave it.
They setup an automated email campaign that prompted people to buy the contents of the wishlist with a promotional code
Went from < 1% buying their wishlist items to +16%
Example of A/B TestWhere did we get the idea?
Test and MeasureThings you can A/B test:
Headline verbiageColors
Call to Actions / ButtonsProduct photosPrice (up sell)
Amount of copy (long or short?)Video?
TestimonialsTrust symbols
Problems...In the last example, the client told us their average dollar
sale was $86.
All of their marketing, acquisition cost tracking, ad spend, etc. were based on that number.
They were wrong, it was $27. Lesson 1: Test your assumptions.
Lesson 2: Put a $ value on every test.
Multivariate Testing
Localization
Benefits
Audience
Headline
Banner
Button / CTA
Trust Seals
After finding out what combination converted the
best.
A conversion was measured initially as someone that got
past the first page.
Then as someone that applied for a loan....
The cost to acquire a new customer was greater than the commission they were
paid.
After an investment of over $20,000 they scrapped the
project.
They were ecstatic to finally figure out why it wasn’t
working.
In the past they had spent over $450,000 with no clue
as to what was wrong.
Tools for testing
Tools for measuring
Recap the riddle3 women are having dinner and the bill comes to $30. Each of them puts in $10. 3 x $10 = $30
As the waiter rings them up, he realizes it should have totaled $25.
The waiter starts to take the $5 back to the women but realizes that the can’t split the $5 over 3 people so he pockets $2 as a tip and gives each woman back $1. Therefore, now they have paid $9 each.
$9 x 3 women = $27, now add in the $2 the waiter kept and that totals $29.
So, what happened to the missing dollar?
The answerThere is no missing dollar...
1. Restaurant kept $25 for the bill2. Waiter kept $2 for his tip3. Customers got back $3 in change
$25+$2+$3 = $30$9 x 3 people = $27
Lesson #1Don’t get caught up in the wrong
details.
Focus on the end goal and figure out what gets you there.
Lesson #2Don’t expect one change to create
massive results.
It is usually a series of changes over time that creates incremental
increases.
Lesson #3If you aren’t sure what to test.
Survey your existing visitors to figure out where they get stuck.
Lesson #4Wait until you have enough statistically relevant data.
Don’t cut your test short, wait until you have enough of a delta to know.
Lesson #5
Most of the time continual tweaking is better than redesigning.
Arif Gangji
Direct: 720-230-9410Email: [email protected]: www.neonrain.com
@arifgan
www.linkedin.com/in/arifgangji