To succeed in today's rapidly evolving marketing landscape, you need to understand how to collect, analyze, and leverage the massive and varied amount of data available. A system of data analysis, usable by data novices and ninjas alike, can unlock your campaigns’ performance potential.
Hear from StrongView’s Senior Strategist, Catherine Magoffin, as she lays out a step-by-step, soup to nuts process for data analysis, focused on digital marketing performance.
Key Topics
* Why it is so important to begin utilizing your customer data, today
* 11 Steps for harnessing your customer data into action
* Real life examples of success from Cooking.com and Redfin
1. Proprietary and Confidential | 1
HEADLINE EXAMPLE
June 19, 2014
11 Steps to Analyze Data for
Campaign Performance
2. Proprietary and Confidential | 2
Welcome!
Today’s Topic:
11 Steps to Analyze Data
for Campaign Performance
Presenter:
Catherine Magoffin, Sr.
Strategist and Team Lead at
StrongView
3. Proprietary and Confidential | 3
Today’s Agenda
• Using Data to Drive Contextual, Present Tense
Marketing Experiences
• Review of the 11 Step Methodology for Data
Analysis
• How Data Translates into Contextual Consumer
Experiences and Marketing Results
6. Proprietary and Confidential | 6
Present Tense Marketing Pillars of Success
1) Acquisition & Revenue
2) Context Awareness
3) Data
4) Efficiency
5) Channel Integration
7. Proprietary and Confidential | 7
Beyond Lifecycle = Present Tense Marketing
Present Tense
Marketing
Single
Channel Multi - Channel Cross -Channel
Evolving the Dialog to the Constantly Connected Consumer
11. Proprietary and Confidential | 11
1. Define the question
2. Define the ideal data set
3. Define what you can access
4. Obtain the data
5. Clean the data
6. Conduct exploratory data analysis
7. Deploy statistical/predictive modeling
8. Interpret results
9. Challenge results
10.Document results and recommendations
11.Outline ongoing data analysis plans
The 11 Steps
12. Proprietary and Confidential | 12
Objective: Clearly specify the general and specific question
you need to answer. This is the MOST IMPORTANT STEP.
Step 1: Define the question
13. Proprietary and Confidential | 13
Any question may be a good question . . .
If it supports your business objectives and program
optimization goals. Think about questions relating to:
Channel Engagement
Device & OS
Activity
Location
Time
Demo-Socio-Psycho-Graphic
Purchase History
Lifecycle Stage
Content Preferences
Permissions
Source
Loyalty levels
14. Proprietary and Confidential | 14
Step 2: Define the ideal data set
Assuming you have access to anything and everything,
define the ideal data set to answer the question.
15. Proprietary and Confidential | 15
Step 3: Define what you can access
Realizing you may not have access to every data point
desired, what can you get? Think about where it resides,
how you can get it and how you can consume it.
18. Proprietary and Confidential | 18
Step 5: Clean the data
Manipulate the data to be usable in your analysis tools.
Remember to keep a clean copy of the original data you
obtained and to describe how you changed it in writing.
19. Proprietary and Confidential | 19
Step 6: Explore the data
Begin to review basics of the data:
• Do you have the data elements needed to
answer the question?
• Is it accessible by key segments and attributes,
such as:
• Program response
• Specific timeframe
• Brand or product category
• Region
• Past purchase or Loyalty Level
• Source
20. Proprietary and Confidential | 20
Step 7: Deploy statistical/predictive modeling
Once you have a basic understanding of the data set, begin
to describe the process, relationship or trends the data is
revealing. What story is it telling? Where necessary, apply
statistical modeling techniques to better assimilate the data.
21. Proprietary and Confidential | 21
Step 8: Interpret results
Once you understand the data model or relationship, what
does it tell you about the broader question? Can you
answer the question now? How does the data answer the
question?
22. Proprietary and Confidential | 22
Step 9: Challenge results
Before presenting the results to stakeholders, have a data
hackathon of sorts -- try to poke holes in the data and your
analysis. Do this yourself and have other colleagues
provide their input and challenge the results.
23. Proprietary and Confidential | 23
Step 10: Document results & recommendations
Finally, present your results, interpretation of the data and
recommendations to key stakeholders. Decide on next
steps and a plan of action.
24. Proprietary and Confidential | 24
Step 11: Document your process
Make sure someone else can come back and consistently
replicate the process. Document all steps, save all files and
make them available for future reference.
25. Proprietary and Confidential | 25
Headline Example
Proprietary and Confidential
Examples of
Data Driving Success
26. Proprietary and Confidential | 26
Analysis and Insight
Analyzing purchase behavior, demographics, location,
interests, buyer scoring and other dimensions to assess the
impact on purchases.
32. Proprietary and Confidential | 32
1. Define the question
2. Define the ideal data set
3. Define what you can access
4. Obtain the data
5. Clean the data
6. Conduct exploratory data analysis
7. Deploy statistical/predictive modeling
8. Interpret results
9. Challenge results
10.Document results and recommendations
11.Outline ongoing data analysis plans
The 11 Steps Recap
34. Proprietary and Confidential | 34
Questions?
• Go to www.strongview.com
• Whitepapers
• Research
• Case Studies
• Webinars
• Expert Advice & Blogs
• Twitter: @strongview
• Facebook.com/strongview
Catherine Magoffin
Sr. Strategist and Team Lead
cmagoffin@strongview.com
650-226-6826
Hinweis der Redaktion
Goal
Bring it all together with the help of a specific example
Sample Narrative
Here’s an example of how the breadth and depth of data we talked about could influence how a marketer engages a customer. We start with a traditional approach where we target a consumer based on their profile data and some past purchase history.
So, We have a fictitious consumer we’re calling “Pete”, who is a 35 year old male from Green Bay Wisconsin who recently bought a washer and dryer. Based on this limited set of data we have on Pete, we might conclude that he is a new home owner whose next purchase might be a refrigerator, and so I might target him with an offer for that appliance.
But if I were able to supplement Pete’s profile data with his web browsing and search history as well as “Likes” from his Facebook profile, that would help paint a more complete picture of Pete. In this case, I would learn that Pete is a Packers fan, and that he’s been pricing comparison shopping 60 inch Samsung televisions. If I were to also factor in the results of last week’s play off games and the weather forecast for the week ahead that shows storms heading into Green Bay, I’d conclude that Pete is in a prime candidate to buy a TV now, and that the best offer would be Free Delivery and Setup of that new TV he wants so that he doesn’t have to deal with the storm that’s coming.
This is the power of Present Tense Marketing.
Questions for Audience
Can you envision a scenario like this being effective with your audiences?
In response to how consumers engage with brands today, marketers need to engage audiences in the context of their current need or situation, which we call “state”. Our belief is that If you as a marketer understand and engage your audience in the context of their current state, then you’ll be more relevant and successful with them. Otherwise, it’s likely you’ll be ignored.
StrongView has developed a methodology call PTM – which reflects the idea that marketers need better insight on the needs and intentions of their customers at specific moments in time, and they need to be able to take action in the context of those insights.
NOTE: As you do this, more questions will likely arise that may require you to access more data. Go back to Step 2 (Define) and continue to refine the ideal data set.
ANOTHER IMAGE?
Demographics (Age, Income, Net Worth, Household Composition, Occupation, Ethnicity, Religion, Voting)
Residence (Location, County Size, Home Type, Home Value, Length of Residence)
Purchase Behavior (Product Categories, Frequent Purchases, Lifestyle Segments)
Interests (Hobbies, Music, Reading, Sports, Travel, Donor, Financial, Diet, Collectibles, Ownership, Shopping)
Buyer Score—A measure of purchasing activity and capacity
Resource putting together buyers and sellers. Have been able to leverage profile data to serve more personal info and proactive – allowing customers / sellers to do price tests – Price Whisperers. Post-sale communication using weather in area – do need to clean gutters?
This is the second email in the series – it’s very powerful because it’s the only way we have for users to let us know if they’re buying and/or selling. When they click we send them to dedicated landing pages with our offerings, and we’re able to store that click for future targeting.