Digital and Data Driven Marketing.
Week 1: Developments in Digital and Direct Marketing, and Customer Acquisition, Retention and Loyalty Online (13 July 20)
E-commerce: No creative process is truly complete until it manifests a tangible reality. Whether your idea is an action or a physical creation, bringing it to life will likely involve the hard work of iteration, testing, and refinement.
E-Commerce Marketing Communications: Contact – Interact – Transact – Relate
Consumer Platforms and Trade
- Social selling
- Multi-channel e-tailing
- M-commerce
- Social media
- Augmented reality
- Location based marketing
In the See, Think, Do, Care model developed by Avinash Kaushik (2015), the customer journey has four stages of consideration and intent:
- See: looking for information.
- Think:thinking about making a purchase.
- Do:ready to make a purchase.
- Care:has made multiple purchases and is a loyal customer.
Marketing strategies for customer retention (at the Care stage) often include customer loyalty and brand loyalty initiatives. Here are some examples:
- Brand love: What is the level of emotional connection with the target consumers?
- Value for money:Are the prices of the product competitive, or is it overpriced when compared to others?
- Word of mouth (WOM): Does the customer recommend your product to family and friends?
- Repeat purchase: Is satisfactory customer service given along with incentives to buy again?
Using data to inform each stage of the journey: Retargeting – Lookalikes – Customer Match
Week 2: Digital Consumer Behaviour and Database Marketing (20 July 20)
Google’s Sridhar Ramaswamy (2015) describes micro moments as follows: ‘Micro-moments occur when people reflexively turn to a device—increasingly a smartphone—to act on a need to learn something, do something, discover something, watch something, or buy something. They are intent-rich moments when decisions are made and preferences shaped.’
‘Database marketing is the use of customer databases to enhance marketing productivity through more effective acquisition, retention, and development of customers’ (Coussement et al., 2016, p.1).
Marketing theorists such as John Egan (2009) suggest that database marketing can be viewed as a technological manifestation of relationship marketing. It is often confused or merged with customer relationship management (CRM).
According to Zikopoulos et al. (2013), over 2.5 quintillion bytes of data is created every day. Customer data includes purchases, web visits, mobile app interactions, social media posts, ‘liked’ ads and call-centre charts.
Databases enable marketers to manage this big data and to use advanced analytics to help them draw insights for their strategies.
Gandomi and Haider (2015) describe eight characteristics of big data.
- Volume
- Value
- Veracity
- Visualization
- Variety
- Velocity
- Viscosity
- Virality
A marketing database contains six types of consumer data:
- who:contact details, status, whether a customer or prospect, and credit status
- how much: price of items, order value and amount of returns
- what:whether an order or enquiry, items ordered and product category
- when:order history and last order (recency, frequency and value – RFV)
- where:sales channels, branch and media
- why: promotions responded to and messages.
Segmenting a customer base and personalizing communications are core principles of direct and digital marketing. They are underpinned by the database.
The primary aim of segmentation is to target consumers individually through automation (Gowens, 2017).
A customer database can be segmented by:
- consumer data, including name and salutation
- behavioural data, including items viewed and shared
- transactional data, including items, time and quantity purchased.
Segmentation is often achieved through approaches such as chi-squared automatic interaction detection (CHAID) analysis. CHAID is a tool used to discover the relationship between variables. CHAID analysis builds a predictive model, or tree, to help determine how variables best merge to explain the outcome in the given dependent variable (Thomas and Housden, 2017).
Despite the availability of these segmentation techniques, 62% of UK marketers are not personalizing their communications (DBM, 2017).
Creating a marketing database involves four key operational phases:
- Database design:identify the data you need and how you can get it.
- Data collection:identify where to collect the data and whether you should incentivise.
- Data mining:analyze what the data is telling you.
- Data cleansing:check that data is current and complies with data protection regulations relating to consent (opt-in and opt-out).
Once a database is created, you can continue collecting data by:
- carrying out customer satisfaction and purchase surveys
- prompting (incentivising) data updates
- capturing transaction and customer contact data.
Use data cleansing to remove:
- customers who want to unsubscribe
- unprofitable customers
- customers who have registered with services such as the telephone and mailing preference services (TPS and MPS)
- duplicate customer records (known as deduplication).
‘The information revolution in the past couple of decades has caused a proliferation of customer databases, often leading to injudicious applications of direct marketing techniques, canvassing the market with ineffective sales pitches, increasing consumer resistance to “junk mail” and telemarketing’ (Kamakura et al., 2003).
This is a warning to database marketers.
Other potential challenges include:
- customer willingness to provide data
- perceptions of lack of privacy
- employee willingness to share data
- the existence of multiple company databases
- the risks involved with predictive analysis
- security issues and potential media coverage
- the lack of human interpretation
- ageing data
- the assumption that past data is an indicator of future trends.
-
- Start with your data
- Enhance and enrich first-party data
- Segment your customer groups
- Create a rich picture of these customers on and offline
- Deploy your targeted cross-channel campaigns
- Measure the effectiveness of your campaigns
- of adequate size: the financial return has the potential to meet your target
- measurable: you can identify segmentation variables that are related to the product purchase and develop a descriptive profile of the segment using a combination of these
- accessible: you can reach the segment in an efficient and cost-effective manner
- a distinct subset of consumers: the group of consumers is distinguishable from other sets, with no overlaps
- compatible: the segment has the ability to purchase.Week 3: Digital Customer Profiling Techniques (27 July 20)Customer profiling is a way to create a portrait of your customers to help you make design decisions concerning your service. Your customers are broken down into groups of customers sharing similar goals and characteristics (Experience UX, n.d.).
Segmentation is the process of selecting and isolating people or companies within a market, based on shared similar characteristics (Thomas and Housden, 2017). It is about understanding customers at an aggregate level.
In order for segmentation to be successful, a segment must be:
Segmentation variables:
- Geographic
- Socio-demographic
- Behavioural
- Psychographic
- Benefits sought
- Geo-demographic
Profiling allows you to delve further into the lifestyles of selected customer segments.
Each segment is given a (typical) representative, with a photo, a name and a description. A small group of customer profiles, or personas, is then used to make key design decisions (Experience UX, n.d.).
6 Steps:
In business-to-business (B2B) profiling, customers can be identified at both the organization and individual employee level.
Common profiling criteria include:
- organization: location, size (number of employees), spend, industry, reputation, structure and payment record
- product specific:number of vehicles or PCs and type of CRM system
- individual: role, location, past behaviour and relationships.
Data mining refers to the automated process of identifying patterns in data that can lead to economic advantage (Witten et al., 2016). It has a broad application across the sciences, law, education, politics, social issues and, increasingly, business.
Data modelling describes a series of mining techniques that are used to identify patterns in pre-existing data sets. These techniques can be used for a wide range of tasks, such as determining preferred customer channels, identifying segments and isolating the most profitable groups. They are performed by software but require human interpretation.
Cluster analysis
Cluster analysis is an exploratory data analysis tool for organizing observed data into meaningful taxonomies, groups or clusters (Stone, 1996). It is an ideal tool for database segmentation.
The cluster analysis process involves:
- identifying the variables you will use to distinguish individuals
- creating clusters using these variables
- running additional tests to see what shared characteristics are within each cluster
- using outcome data as a source of future targeting and personalisation decisions.
FRAC analysis uses a points-based system where customers are rated on the following:
- frequency:the number of purchases within the last 12 months
- recency: purchases made, for example, in the last three, six, nine or 12 months
- amount: the average spend per purchase
- category:the product range purchased from.
FRAC analysis can be used to rank customers and to shape retention and dissolution activity. It can also factor in measuring predicted lifetime value (LTV) and allowable spend per customer (Tapp et al., 2014).
Regression analysis
In marketing terms, regression analysis is a technique used to identify the influence that a range of independent variables may have on another, dependent, variable. It is the most tried-and-trusted technique used by database marketers (Tapp et al., 2014).
CHAID analysis is a decision tree technique widely used in direct marketing. It works by repeatedly splitting segments into smaller segments until no more statistically significant splits are possible. Unlike clustering, you select the order in which criteria are used to segment, and each segment is clearly distinct from others.
CHAID analysis provides a clear picture of the type of person most likely to purchase products and services, based on factual purchase history, geo-demographics and lifestyle attributes.
