personalization best practices for engaging online experiences

Post on 09-May-2015

3.499 Views

Category:

Business

2 Downloads

Preview:

Click to see full reader

DESCRIPTION

Slides from an American Marketing Association (AMA) webinar re: how to implement online personalization techniques - pitfalls, opportunities, best practices.

TRANSCRIPT

Personalization Best Practices for Delivering

an Engaging Online Customer Experience

Carlos Carvajal, Sr. Director of Marketing, Baynote

April 28th, 2010

Agenda

• Personalization 101

• Common Pitfalls

• Always Right Personalization

• Real-World Use Cases & Results

What is Personalization?

EXPLICIT IMPLICIT

• Profile-based

• Segment-based

• Rules-driven

• Behavioral-based

• Context-driven

• Collaborative Filtering

Personalization is a strategy for delivering a customer

experience that is always personal, relevant and convenient

Many Ways to Personalize… Implicit

Product

Recommendations

Content

Recommendations

Many Ways to Personalize… Explicit

CustomizationRole-based

Many Ways to Personalize… Channels & Content

Personalization Benefits

Increase:

• Conversions

• Engagement

• Revenue

• Customer Insights

• Customer Lifetime Value

Common Personalization Pitfalls

Four Common Personalization Pitfalls

1.Must Build a Comprehensive

Customer Profile First

2.Create Rules to Target Content

Based on Customer Preferences

3.“Ultimate Personalization!”

4.Personalization is Really Hard to

Get Right

Pitfall #1: Must Build A Comprehensive Profile First…

Father

Uncle

Husband

CEO

Brother

Engineer

What to recommend???

Pitfall #2: Create Rules to Target Content…

Pitfall #3: “Ultimate Personalization!”

Pitfall #4: Personalization is Really Hard to Get Right

BI, Analytics & MV Testing

• Too much noisy data

• Rear-view mirror

• Guess work

Re-design

Re-launch

Re-platform

Annually

Quarterly

Monthly

Tuning, Merchandizing & Editing

• Manual & costly

• Top 5% most popular only

• Expert bias

Development & QA

• Too many projects

• Conflicting priorities

• Expensive

Approval & Deployment

• New issues with priority

• Missed deadlines

• Out-of-sync with market

Too Slow, Too Expensive

Personalization in 2001 – Learn From the Past

Source: ‘Web site personalization’, IBM 2001

Always Right Personalization

VS.

Always Right Personalization

“By understanding who you are and your shopping

history… I can accurately predict which products you are

currently interested in.”

“What are you looking for?”

“Many customers who came before you liked this

beginner snowboard…”

Collective Intelligence Powers Always Right Personalization

Invisible CrowdsActions > Words

Whole > Sum of parts

Real-time feedback

Crowds are freeCrowds > Expert

Crowds reduces cost

7 deadly biases

Aligned incentives?

Sell 1 of manyNiche segments

Micro projects

Real-time trends

But how?

Combination of Social and Cognitive Science

Respond to Changing Customer Needs in Real-Time

BI, Analytics & MV Testing

• Too much noisy data

• Rear-view mirror

• Guess work

Re-design

Re-launch

Re-platform

Annually

Quarterly

Monthly

Tuning & Merchandizing

• Manual & costly

• Top 5% most popular only

• Expert bias

Development & QA

• Too many projects

• Conflicting priorities

• Expensive

Approval & Deployment

• New issues with priority

• Missed deadlines

• Out-of-sync with market

Automatically Observes

Emergent Patterns

Use Collective Intelligence

Little IT Involvement

Real-time Automatic Adjustments

Three Principles for Achieving Always Right Personalization

1.Focus on Context First

2.Understand Customer Intent

Through Engagement

3.Like-Minded Peers Know What

They Like Best

Principle #1: Focus on Context First

What to recommend

?CONTEXT DRIVES RELEVANCE

Father

Uncle

Husband

CEO

Brother

Engineer

Principle #2: Understand Intent Through Engagement

Principle #3: Like-Minded Peers Know What They Like Best

Putting it All Together…

…page refers, links, entry trails,

queries, page sizes, mouse movement,

negative experience, virtual bookmarks,

time spent, exit trails…

Real-World Use Cases & Results

Use Case: Personalized Landing Pages

Unknown context

Known context

Use Case: Personalized Search

The crowd knows “insight” is really the LG Incite phone.

Use Case: Dynamic Navigation

• Real-time as trends emerge

• Based on aggregated visitor

behavior

• Brings content to the

navigation level

Popular content in category

Use Case: E-Mail Personalization

4x

Conversion Rate

Promotional EmailTransactional Email

Use Case: Video Recommendations

21% Increase in

Video Starts

Use Case: Recommendations Across Content Types

Most Popular, Related, Content Video, Email…

Use Case: Social Personalization

What I‟m cooking

today

You Might Like

Items shared by friends drive recommendations based on your preferences

Use Case: Identify Customer Trends

Identify Trends

Before They Happen

Find Gaps and Optimize

Product Mix

Use Case: Merchandising & Targeting Controls

Select Action

Select

Product

Select

where

product

should be

shown

Personal Profiles Serve as Filters and Constraints

Personalized store

Personalization Best Practices

Personalized Contextual Recommendations

Contextual Recommendations

Known

Unknown

Known

Personal RecommendationsMost Popular Products

Unknown Profile

Context

Real World Results: Bluefly

Profile• $84M High fashion discount eTailer• 50K SKUS, 350 brands

Results:300% lift in sales in Q4 „09

“By making recommendations more relevant from a time perspective, we knew we would get a lift, but we had no idea it would be this much"

– Marty Keane, SVP ecommerce

Challenges• Low inventory per item• Constant turnover of product• Needed real-time solution

Real World Results: TurboTax

Profile

• 14M taxpayers

• Award-winning support site

Results:• 35% increase in site search

conversion rates within weeks

• 73% increase within a year

“The results were so overwhelmingly positive that it became a no-brainer.”

Chris Jones, Web Analytics Manager

Challenges• Taxpayers unable to find answers• Manually hard-coding content

recommendations in results• Rapidly changing content &

seasonality

Baynote in a Nutshell

Company• Pioneered real-time recommendations

using collective intelligence in 2005

• 200+ web sites use Baynote today

Technology• Collective intelligence foundation

• Breakthrough UseRank® patents

Solutions• Real-time personalization, targeting and

recommendations

• SEO and SEM – peer-driven

• Search optimization – re-ranking by

engagement

• Email & video recommendations

• Insights - Merchandising, Testing,

Reporting

Key Takeaways

1. Avoid Common Pitfalls• Build a comprehensive customer profile first

• Create rules to target content…

• Personalization is really hard to get right

2. Best Practices for Success• Understand customer intent thru engagement

• Use Collective Intelligence to respond to

changing customer needs in real-time

• Personal profiles serve as filters & constraints

3. Start Simple…• Identify high-value personalization use case:

landing page, e-mail, product detail page

• Define clear success metrics and plan for

results within 8-12 weeks

• Use customer insights, run tests, adapt

Thank You for Your Participation!

• E-mail: carlos@baynote.com

• Twitter: @Baynote

Contact Information

Continue the Conversation on Twitter: #AMABaynote

• www.baynote.com

• www.baynote.com/blog

To Learn More About Baynote

top related