turning digital marketers into data scientists

6
TURNING DIGITAL MARKETERS INTO DATA SCIENTISTS Creating Custom Audiences for Paid Search

Upload: adlucent

Post on 18-Feb-2017

30 views

Category:

Marketing


1 download

TRANSCRIPT

Page 1: Turning Digital Marketers into Data Scientists

1

TURNING DIGITAL MARKETERS INTO DATA SCIENTISTS Creating Custom Audiences for Paid Search

Page 2: Turning Digital Marketers into Data Scientists

Consumer behavior is changing and it’s having a huge impact on the retail industry.

Storefront sales continue to decline while retail sales overall have stagnated.

Shoppers are investing a greater share of their time and wallets into digital, which

grew an impressive 13.4% over the holidays,1 and advertisers have followed suit.

Competition in the digital space is growing and it’s becoming harder to find new

customers and expand reach without simply throwing more money at the problem.

It’s time for advertisers to look inward for a solution as the answer often lies within

the customer data they already own.

By combining the power of intent data with a rich understanding of who our

customers are, we have the opportunity to create more relevant and efficient

advertising. But first, we have to change the way we operate and begin to think more

like data scientists.

Getting the Right Ads to the Right PeopleNearly 70% of online consumers agree that the quality, timing and relevance of

a brand’s message influences their perception of a brand,2 further proving the

importance of customer-level marketing. There is no shortage of data that we can

use to deliver a highly relevant advertising experience for each consumer. In fact, the

volume of information that we collect and store grows every second, but managing

and making sense of this data has become increasingly complex. According to IBM,

80%3 of the data that we collect is dark, meaning it’s unstructured and unused. By

2020, this number will grow to 93%. Our job as marketers is to figure out new ways

to utilize this goldmine of information to identify and capture the right customers for

our business while creating the best brand experience possible. And this is going

to be a bigger priority for digital marketers in the year ahead, with 70% stating that

cross-device audience recognition will get the greatest proportion of their attention in

2016.4

Fortunately, you don’t have to be a professional data scientist to build more

personalization into your program. You can start by leveraging the customer data

in your CRM and using Google’s Customer Match tool to extend your reach across

Search, Gmail and YouTube.

Example: TargetTarget uses data from their baby registry to deliver relevant ads and offers, such as this first birthday mailer, to parents.

“EVERYONE IS NOT YOUR CUSTOMER”

-Seth Godin

% of Consumers

70Timing is Everything 70% of consumers say quality, timing and relevance of a brand’s message influences their perception of a brand– Google

Page 3: Turning Digital Marketers into Data Scientists

Google’s Customer MatchHistorically we’ve only been able to use keywords and intent data to target shoppers on

search networks, but thanks to recent changes from Google, we’re now able to serve

ads to individuals we have CRM data for and others who share similar characteristics.

This is the first time Google has allowed advertisers to use their own first party data to

reach customers in Adwords using customer email addresses. Customer Match provides

advertisers with the ability to upload a list of email addresses directly into AdWords

which are matched to signed-in users on Google in a secure and privacy-safe way. The

email addresses you upload do not have to be Gmail-only, as Google will try to match

non-Gmail addresses to specific users. Adlucent clients have seen average match rates

of 54%. Once the email uploading and matching is complete, advertisers can target, or

exclude, matched individuals across Gmail, Search, and YouTube. In addition, you can

generate Similar Audiences to reach new customers on YouTube and Gmail who are likely

to be interested in your products and services, thus expanding your reach. With billions of

monthly searches on Google, and millions of Gmail and YouTube users, the addressable

market is huge. This is a great first step toward more personalized advertising.

How Customer Match WorksLet’s say we want to use Customer Match to advertise to customers who haven’t made

a purchase recently. We can create a “haven’t interacted with your brand recently” list

and add these customers to our other one-time buyers list. We’ll then upload the email

addresses for these lists into AdWords where we’ll use this data to adjust bidding and can

serve specific messages such as “come back, we miss you!”.

Ways to Use Customer Match• Reactivate lapsed customers: Target customers you know, but who haven’t

purchased in a while

• Prospect new customers: Exclude customers you know and only target new ones

• Cross sell existing customers: When you know what customers have already

purchased you can suggest something complimentary

• Target similar customers: When you know what your best customers look like,

Google helps you find similar groups and serve them ads on YouTube and Gmail

• A/B Testing: Customer Match allows for easy A/B testing as you can split out a

group of email addresses that fall into the same segment, making the quality of

your control and test groups fairly balanced. You can test different bid amounts on

these groups

How to Measure Performance We recommend using the following KPIs when running Customer Match: RPC (Revenue

Per Click), CVR (Conversion Rate) and CPC (Cost Per Click). A higher RPC means

CVR and/or AOV (Average Order Value) are also high, which is why we see RPC as an

important metric to focus on.

Customer MatchCustomer Match allows advertisers to send targeted messages to pre-set groups of customers

Page 4: Turning Digital Marketers into Data Scientists

Customer Match ResultsWe launched Customer Match for our client Jarden Corporation, a company that

produces and promotes over 120 consumer brands sold globally such as Crock

Pot, Food Saver, Mr. Coffee, Sunbeam, and many other household names. We

tested one of their well-known products with the goal of gaining more efficiency

in brand terms. Through Customer Match, we targeted previous customers on the

Google Search Network and applied a higher bid modifier in order to increase our

visibility with these individuals. The result was an increase of 67% in RPC, and 102%

increase in CVR, with a decrease of 6% in CPC. By focusing ads on those who are

already familiar with the brand, we were able to sell more products with a smaller ad

investment.

Taking Customer Match to the Next LevelWhile uploading email addresses directly into Adwords may help you gain efficiency,

segmenting your data before it gets uploaded into Adwords will help ensure you’re

delivering the right message to the right people. When you take the information in

your CRM (name, email, SKU purchased, date of purchase, value of purchase, etc.)

and combine it with transactional data from your search program (past purchases

made, ad type clicked, categories of interest, historical spend and more), you get a

more complete picture of what your customers look like. This information can be

used to create high value segments to target. For example, you may want to group

customers who have previously purchased football related sporting equipment and

serve them different offers than those who purchased baseball equipment. Your

products and services are unique to your business, so start by thinking about the

various audiences you would like to reach and use your CRM and transactional data

to find the right members for those groups.

How Adlucent Uses Customer MatchAt Adlucent, our Data Science team has been focused on helping our clients deliver

customer-level advertising for three years by leveraging CRM, transactional data,

and 3rd party (demographic, behavioral, lifestyle) data. While most agencies are now

starting to cover basics such as using email addresses in Customer Match, Adlucent

has been at the forefront of this trend, going beyond email addresses to also include

additional data sources.

Here is an example of some of the data we analyze:

CRM

Historical transaction

Historical spend

Social media usage

Interests

Demographic

Hobbies

Behavioral

Lifestyle

In-market

RPC CVR

CPCs

Jarden CorporationBy using Customer Match to target customers who were already familiar with their brand, Jarden was able to sell more FoodSaver products at a lower cost.

25%

50%

75%

100%

0%

Name

[email protected]

One time buyer

Purchased SKU 69864184

November 15, 2015

Spent $103.48

Retailer’s CRM Data

Purchased Hayden Metallic V Neck Midi Dress

Converted on mobile PLA

Low value in relation to other customers

Chicago, IL

Transactional Data

+35% +40% +30%

Smart BiddingAdvertisers can adjust their bidding strategy based on what they know about each customer.

Page 5: Turning Digital Marketers into Data Scientists

MethodologyAt Adlucent, we import our client’s full CRM data into Deep Search®, our proprietary analytics platform used to make decisions for digital advertising programs. Once imported, we then merge this data with a client’s transactional data to develop a better understanding of a customer’s shopping patterns. With these insights, we learn more about who these customers are and how to address them more effectively with digital advertising.

Once a retailers CRM data has been imported into our platform, we use machine learning techniques to identify valued customer segments. Customers who fall into these groups are uploaded into AdWords for Customer Match. We can decrease bids for consumers who do not meet specific criteria and increase bids for those who do.

Here are some examples of insights we might use to adjust our bidding strategy:• Those who are already familiar with a retailer’s brand• The device(s) converted on • Types of products based on previous interest • Complementary products

Basic Segmentation + Customer Match in ActionWe applied our advanced segmentation to an IR100 retailer with stores nationwide. Our goal was to use Customer Match to optimize non-brand performance. We selected prominent categories (for example—apparel, shoes, accessories, other) with very unique customer behavior patterns. From there we determined which customers were high or low value based on shopping patterns and behaviors. We targeted these customers across campaigns and significantly increased bids for those we deemed high value customers. By taking this approach, we saw a 34% increase in RPC and a 64% increase in CVR for our high value segments.

We also did a test to see whether customers who bought from a specific category in the past would convert better from that same category’s campaigns—for example, showing more apparel ads to those who already made an apparel purchase. The results of the test showed that ads for products in complementary categories were more effective than same category ads for this retailer. As a result, we adjusted our strategy and increased bids for cross sell and complementary categories. We saw AOV increase 18% and an 8% increase in Return on Ad Spend (ROAS).

Your product set is unique, so you may have dramatically different results. Make sure to test different groups and categories, and to continuously make adjustments to strategy in order to optimize results.

Using Advanced Customer Profiles to Uncover New AudiencesIn addition to utilizing a retailer’s CRM and transactional data, Adlucent can also leverage 3rd party data, such as demographic and behavioral data, to gain a more complete picture of a retailer’s customers. By integrating third party data into our machine learning algorithm, we infer the presence of potential customers everywhere. Utilizing these advanced customer profiles, we can find new similar customers through lookalike modeling and reach them through advanced bidding strategies.

RPC CVR

25%

50%

75%

100%

0%

IR 100 Retailer Adlucent created high value categories for a IR 100 Retailer and targeted customers associated with these categories in Customer Match.

24-34 years old

Female

Purchases beauty and cosmetic products

AOV $100-$150

5-6 purchases per year

Urban

Domestic traveler

Enjoys tennis

Does not have children

$75,000/yr household income

High value customer

24-34 years old

Female

College graduate

Domestic traveler

Enjoys tennis

Does not have children

$75,000/yr household income

3rd Party Data

Page 6: Turning Digital Marketers into Data Scientists

Applications to Other Business ChannelsArmed with these additional insights about customers, marketers can apply them to other channels in their business. Much like they are used within paid search, these insights will help determine which geographic areas hold the most potential, and which do not. Here are some examples:• Improve budget allocation across the entire marketing mix• Refine email strategies personalized to specific customer profiles, such as loyal

shoppers• Improve catalog distribution lists by reducing catalog mailings to underperforming

households and increasing them in high performing households• Align marketing efforts across all channels by targeting individuals online who

receive offline media such as mailings, billboards, radio, and TV• Use high performing locations to identify potential new locations for brick and mortar

stores• Combine the most popular product data for the relating zip codes to come up with

targeted offers in already existing brick and mortar stores

Turning Digital Marketers into Data Scientists The opportunity to leverage even more data within paid search puts us on the cusp of truly being able to humanize advertising. Combining the power of intent data with a rich understanding of your customers provides us with new opportunities to create the most powerful type of paid search advertising yet.

We are no longer blind to the consumer on the other side of the keyboard, seeing only search terms to navigate by. Instead we find ourselves in a position to better serve our customers’ needs by understanding more about who they are and what they want. This puts us in a position to serve up ads that are more relevant to their individual interests.

Learning how to use data to create much more focused and relevant advertising is critical to creating sustained growth and delivering the best experience for consumers.If you’re interested in learning other ways to leverage your data for more customer-focused advertising, please don’t hesitate to contact us.

About AdlucentAdlucent, a full-service digital marketing and analytics company, helps brands acquire more of the right customer profitably. Adlucent translates vast first and third party data into actionable consumer intelligence, answering questions like “who are my most valuable customers” and “how can I acquire more like them?” These insights are used to connect brands with high value consumers through the right mix of touch points along their path to purchase.

Let’s Chatw: adlucent.com

p: 1.800.788.9152

e. [email protected]

1 eMarketer, Lessons from Holiday Shopping 2015 - What You Need to Know for 2016, Feb 25, 2016

2 Google, Google Brings You Closer to Your Customers in the Moments that Matter, September 27, 20153 Barron’s, IBM’s Rometty Wants You to Know They’re a ‘Cognitive Solutions Cloud Platform Company’, January 6, 20164 IAB Data Center of Excellence, The Outlook for Data 2016: A Snapshot Into Digital Media and the Evolving Role of Audience Insight Research, January 2016