revenue attribution - ted stites, numeric analytics

44
AIM 2010 Revenue Attribution Ted Stites Vice President of Consulting

Upload: joshua-tree-internet-media-llc

Post on 06-May-2015

1.376 views

Category:

Documents


0 download

DESCRIPTION

Attributing revenue to online advertising - from click to lease. A presentation by Ted Stites, Vce President of Consulting at Numeric Analytics to the Apartment Internet Marketing Conference 2010 - aimconf.com.

TRANSCRIPT

Page 1: Revenue Attribution - Ted Stites, Numeric Analytics

AIM 2010

Revenue Attribution

Ted StitesVice President of Consulting

Page 2: Revenue Attribution - Ted Stites, Numeric Analytics

Objective

The purpose of this discussion is to identify the challenges, required data elements and components in answering the question; What is Working?, with Revenue Attribution.

Page 3: Revenue Attribution - Ted Stites, Numeric Analytics

The Tough Questions

• How many leads did each campaign drive?• How much revenue did a particular ILS drive?• Should I be doing Social Media?• What is working?

Page 4: Revenue Attribution - Ted Stites, Numeric Analytics

Common Challenges

• Many are wrestling with these questions.• Similar challenges to retail businesses.• Seems like a simple question, but can have

some complex answers.• Many approaches to answering these

questions. One is applying Revenue Attribution Models.

Page 5: Revenue Attribution - Ted Stites, Numeric Analytics

Sign of the Times

Don't Know

Less than 1%

1%-5%

6%-10%

11%-19%

20%-29%

30% or more

0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50%

0.01

0.01

0.12

0.17

0.24

0.44

1%

1%

21%

22%

20%

17%

16%

TodayIn 5 Years

Source: “Next-Generation Marketing and Measurement,” a commissioned study conducted by Forrester Consulting on behalf of Omniture, June 2009

What percentage of your program budget (excluding personnel/salaries) is spent on online media today? What percentage of your program budget do you predict will be spent on online media five years from now?

Page 6: Revenue Attribution - Ted Stites, Numeric Analytics

6

When & How is Hyper-Complex

@

emailAgent

Viral

Gaming

Social Network

Casts & Txt

MobileCableATM

TV / Mass

Mail / Statements

Retail

IVR

Web

Vending

Lower Touch

Automated Interaction

Transaction Ease

Agnostic Customers

Low Switching Costs

+

Page 7: Revenue Attribution - Ted Stites, Numeric Analytics

7

SeeValue

Need CommitPay / Assess Value

Issue / Need More

The Baby Boomer

TV/Mass Retail Retail Agent

The Millennial

Social Network Viral Web ePay Casts & Txt

Retail

BalanceCost & Ease

TouchFactor

ContentFactor

When & How is Hyper-Complex

Page 8: Revenue Attribution - Ted Stites, Numeric Analytics

Keys to SuccessEngage in cross-channel

measurement and analytics

Identify high-value segments

Deliver targeted offers

2.5X higher revenue growth rates

5X higher customer conversion growth rates

3X higher customer profitability growth rates

1 “Cross-Channel Campaign Management: Next Generation Multichannel Marketing”Aberdeen Group, October 2008

Page 9: Revenue Attribution - Ted Stites, Numeric Analytics

"Never confuse motion with action.“

~ Benjamin Franklin

Yeah… we are doing that

Page 10: Revenue Attribution - Ted Stites, Numeric Analytics

Measuring Value

Marketing effectiveness $1m

Reducing duplicate spend $.2m

Campaign effectiveness $.5m

Customer experience $.3m

Deliverables

Rev Increase

Customer Profit Mgmt

Reduce Cost

Critical Success Factors

25%Return

on Spend

Benefit

Directly impacts NOI

Page 11: Revenue Attribution - Ted Stites, Numeric Analytics

Others that use Revenue Attribution

ResultReduced Paid Search Cost Per Lead by 173% while maintaining conversion rates,

which resulted in over $1 million in annual savings.

Approach

Multiple Attribution Views

Current StateHigh Acquisition Costs Low Measurement Capabilities

Objective

Produce the same with less money

Page 12: Revenue Attribution - Ted Stites, Numeric Analytics

Others that use Revenue Attribution

ResultAllowed for site selection, scoring and prioritization of leads for marketing campaigns.The

highest-ranked leads had a 48% higher response rate than average.

Approach

Graphical density maps of the U.S., scoring table ranking zip codes by propensity to hold higher concentrations of ‘best’ customers created.

Current StateLots of data, with little segmentation Lacked ability to tie segmentation to

financials

Objective

Prioritize site selection and customer prospecting by attempting to locate potential customers and/or areas most likely to hold high numbers of “best” (i.e., high-value) customers.

Page 13: Revenue Attribution - Ted Stites, Numeric Analytics

Others that use Revenue Attribution

ResultThe highest-ranked leads had a 48% higher response rate than average.

Approach

Model was used to score prospects in areas targeted for acquisitions

The model validated successfully across several years and across different geographic markets.

Current StatePossessed past program data Lacked ability to rank customer prospects on

probability

Objective

Client wanted to improve sales conversion by differentiating leads based on customer propensity to accept a wireless service offer.

Page 14: Revenue Attribution - Ted Stites, Numeric Analytics

Lease

Lead

1000

300

What we know…• I know I added 1,000 leads this

week.• I know I wrote 300 leases this week.

What we don’t know…

• Which leases were driven by ILS’s,

Search Marketing, Email, CraigsList?

• Did we touched them multiple

times and if so, how/where?

• How did they interact with my ads?

• How do I make it better?

Page 15: Revenue Attribution - Ted Stites, Numeric Analytics

Lease

Lead

Marketing Channels & Tactics

Page 16: Revenue Attribution - Ted Stites, Numeric Analytics

Lead

Marketing Channels & Tactics

Lease

What’s working?

Page 17: Revenue Attribution - Ted Stites, Numeric Analytics

AIM 2010

Revenue Attribution

A Complex Question…

PlatformsProcesses

People

Page 18: Revenue Attribution - Ted Stites, Numeric Analytics

What does real behavior look like?

Understanding Behavior

Page 19: Revenue Attribution - Ted Stites, Numeric Analytics

SEO SEM ILS Craigslist TV Print Radio Billboard

Lease

Phone Click-Through Email Walk - In

Responses

Lead

Website Mobile Onsite

Advertisements

$

Left Site

Page 20: Revenue Attribution - Ted Stites, Numeric Analytics

What data elements should I consider when building a revenue attribution model?

The Right Data

Page 21: Revenue Attribution - Ted Stites, Numeric Analytics

SEO SEM

Lease

Phone Click-Through

Responses

Lead

Website Onsite

Advertisements

$

Left Site

Data Elements1. Channels2. Touch Order3. Granularity

Page 22: Revenue Attribution - Ted Stites, Numeric Analytics

SEO SEM

Lease

Phone Click-Through

Responses

Lead

Website Onsite

Advertisements

$

Left Site

Data Elements1. Length of

Time

Page 23: Revenue Attribution - Ted Stites, Numeric Analytics

SEO SEM

Lease

Phone Click-Through

Responses

Lead

Website Onsite

Advertisements

$

Left Site

Data Elements1. Date of

Allocation

Page 24: Revenue Attribution - Ted Stites, Numeric Analytics

AIM 2010

The Platforms

Page 25: Revenue Attribution - Ted Stites, Numeric Analytics

SEO SEM ILS Craigslist TV Print Radio Billboard

Lease

Phone Click-Through Email Walk - In

Responses

Lead

Website Mobile Onsite

Advertisements

Ad Serving & Tracking

Analytics

Lead ManagementLease

PropertyManagement

Guest Card

Page 26: Revenue Attribution - Ted Stites, Numeric Analytics

AIM 2010

The Processes

Page 27: Revenue Attribution - Ted Stites, Numeric Analytics

Processes

• Tagging for Analytic Data

• Capturing Lead Data

• Guest Card

• Property Management Data

• 3rd Party Data

• Reporting

• Modeling

Page 28: Revenue Attribution - Ted Stites, Numeric Analytics

Attribution ModelsFirst Touch Last Touch Revenue

First $1,000 $1,000

Last $1,000 $1,000

Linear $500 $500 $1,000

Weighted $250 $750 $1,000

Regression & Media Mix

Page 29: Revenue Attribution - Ted Stites, Numeric Analytics

AIM 2010

The People

Page 30: Revenue Attribution - Ted Stites, Numeric Analytics

Cross Functional Expertise

Data management and segmentation

Technology Integration

Strategy

5%

12%

48%

15%

8%

10%

12%

12%

4%

Source: “Next-Generation Marketing and Measurement,” a commissioned study conducted by Forrester Consulting on behalf of Omniture, June 2009

What skills do you predict will be most critical to your marketing organization over the next five years?

*Top 3 Responses

1st Choice 2nd Choice 3rd Choice

Page 31: Revenue Attribution - Ted Stites, Numeric Analytics

Well Rounded People

Page 32: Revenue Attribution - Ted Stites, Numeric Analytics

AIM 2010

Now that we know what’s working…

How do I make it Better?

Page 33: Revenue Attribution - Ted Stites, Numeric Analytics

Communication

Page 34: Revenue Attribution - Ted Stites, Numeric Analytics

Visited > 1 50%Visited > 3 30%Average Research 66 day

Prior Research

NoResearch

Rent.com18%

Apartments.com32%

RentCarmel.com26%

CraigsList14%

Other10%

Conversion by Channel

Responses

Leads

Leases

255 207 1534 750 1765 962

CraigsEmail ILS Walk-InSEM Phone

1500 1687 5897 3452 5421 4158

3585124485256774

What responses, leads, and leases did each channel

produce?What percent of people are researching via multiple

channels?

How many channels are they visiting and how long are they

researching?

What channel did they visit prior to the converting

channel?

Page 35: Revenue Attribution - Ted Stites, Numeric Analytics

Touches > 1 40%Touches > 3 35%Average Touches 2.5

Responses by TypeLe

ases

OneTouch

More than 2

How did people respond to us?

How many times are we touching them before

converting?

What is the average times we touch a customer?

Page 36: Revenue Attribution - Ted Stites, Numeric Analytics

Cost Per Lease

Average Cost Per Lease$1,900Lead Lease Rate

28%

What is the Cost per Lease

of each Channel?

What is our CPL trending up over

the recent weeks?

What is our average CPL & at what rate are we converting them?

Page 37: Revenue Attribution - Ted Stites, Numeric Analytics

AIM 2010

Where can we go with this?

Tracking

Learning

Modeling

Page 38: Revenue Attribution - Ted Stites, Numeric Analytics

How else?

1. Damage Deposits

2. # of Applications vs. # of Approvals

3. Its more than revenue attribution, it also becomes

Customer Experience

Page 39: Revenue Attribution - Ted Stites, Numeric Analytics

Progression

1. Different Models

2. Establish the foundation to connect the dots

3. Start Simple and add to it based on learning's

Page 40: Revenue Attribution - Ted Stites, Numeric Analytics

Customer Lifecycle Framework

Bad Debt/Churn

0

500

1000

1500

-500

-1000

-1500

-2000

Cum

ulat

ive

Cus

tom

er L

ifetim

e V

alue

In D

olla

rs

Months of Life1 2 3 4 12 13 36 37 38 39 40 41

RevenuesCosts

Market

Acquisition

Fulfill

Cash and Costto Serve

Deposits

RenewalPromos

Market Acquire Fulfill Serve Up Sell

Maintain

Churn

Up sell

Maintain

CustomerCare Costs

*McKinsey

By managing the Customer Lifecycle indicated a company can increase lifetime value and consequently EBITDA by up to 20%*.

Page 41: Revenue Attribution - Ted Stites, Numeric Analytics

How do you leverage segmentation?Relative Customer Value

Cost to Serve

Rev

enu

e

Low

Lo

w

High

Hig

h

Valuable-Low TouchCustomers

Valuable-High Touch Customers

Average Customers

High Touch Customers

Most Profitable

Least Profitable

Develop an understanding of which customers to focus retention efforts

Retain

Create initiatives designed to migrate low revenue customers toward being valuable customers

Grow

Page 42: Revenue Attribution - Ted Stites, Numeric Analytics

AIM 2010

Yeah, but

Page 43: Revenue Attribution - Ted Stites, Numeric Analytics

Have the ability to Act

• Budgets

• Automation

Page 44: Revenue Attribution - Ted Stites, Numeric Analytics

In Closing…

1. It can be complex, but don’t let that hold you back. There are big

opportunities to be had.

2. Identify the data elements and attribution model(s) you want to apply.

3. Assess your current state and capabilities, then create a competency and

progression plan.

4. Establish the foundation with a complete data capture strategy.

5. Clearly define the questions that you want to answer, don’t try to measure

everything.

Ted [email protected]