Predictive analytics in Marketing: Segmentation and beyond – Why should I care?
Your marketing ROI will have been multiplied (usually by 750 %) – so it worth a try. Predictive Analytics in Marketing is not an utopia – it is the reality for DnA Clients.
Predictive Analytics in Marketing – Behavioural segmentation process example
DnA gradual introduction approach
All the methods used are constantly improving, the data you have is different than others have, the DnA algorithms are always and continuously evolving – so accordingly any DnA Analytics Process is going to be introduced gradually so we developed our own step by step approach. You do not have to start a multi-million dollar project lasting years full of scope creeps; we at DnA begin with small steps and demonstrate the results within days – then we can move forward and gradually built a full winning marketing based empire within months.
Customer behaviour seems just an outcome of a complex process: primary motivations, experiential motivations, behaviours and results are the main stages.
Your Company Results (typically financial) are caused by behaviours (usually some kind of transactions, purchases and marketing communication responses), which are caused by one or both (primary and experiential) motivations.
Some motivational causes (searching, need, arousal, etc.) can exist no or just limited brand interaction. On the other hand primary motivations (price valuation, attitudes about lifestyle, tastes and preferences, etc.) are generally psycho-graphic and hide under the surface so not really seen. Experiential motivations tend to have brand interaction and are another motivator to additional behaviours that ultimately cause (financial) results. These motivations are things like loyalty, engagement, satisfaction but be warned engagement is not a behaviour this is an experiential cause – result of interactions with the brand & co.
How Prescriptive Marketing Analytics helps to set marketing & customer strategy: marketing is customer-centric
DnA process starts with business understanding: a practical discussion about the company’s current challenges and we actively assist to formulate the ultimate business question on a predictive marketing analytics use case. Do not worry, at DnA we have a full Analytics Due diligence Method (Analytical Audit) to identify and form these opportunities and identify gaps and traps within the optimal trajectory. Moreover we have established processes to improve current business practices beyond Marketing; Sales, HR, Product Development, Customer Service, etc. Practically every business area and practice can be improved by Prescriptive Analytics – we know this for sure. After establishing goals, a strategy needs to be in place to reach those goals. A very different segmentation should result if the strategy is about market share as opposed to a strategy about net margin. DnA Bootloop Analytic Engine reverse engineers the best marketing strategy approach – but that is not for the faint hearted: that is applied prescriptive marketing, and specifically operational after collecting a lot of data and applied predictive analytic approach in practice [the engine needs market-reactive attributes, eg. how the market / customers reacted to the actions you made…] so let us leave it for later.
The successful marketing strategy discussion includes customer behaviour, so state at least one of your relevant business questions related to:
- What is the behaviour we are trying to understand?
- What is the mindset in a customer’s mind?
- What incentive are we employing?
Any good segmentation solution should tie together customer behaviour and marketing strategy therefore starts with business question. This is what DnA can focus on along the process.
- Segmentation provides insights into marketing research, marketing strategy, marketing communications and marketing economics.
- Segmentation is about what is important to your consumers, not what is important to your firm.
- Each segment required to have its own story rationale for why it exists and treated as a segment.
- There should be a different strategy levelled at each segment, otherwise there is no point in being a segment.
Applied Predictive Analytics in Marketing – DnA Process by large
Collect appropriate (behavioural related) data
This tends to be generally around transactions (purchases) and marketing communication related responses and act as main dimensions of behavioural segmentation
We want to know how many times a customer purchased, how much each time, what products were purchased, what categories each product purchased belonged to, etc.
DnA Predictive Marketing Process automatically generates additional profiling variables including net margin on each purchase, cost of goods sold, number of transactions over time, number of units and if any discounts were applied to these transactions, etc. Of course collecting vehicles / channels and related info, store purchases, coupons used, website visits are essential to success. All of this data surrounding transactions and responses is the basis of customer behaviour finding segments of category, preferred communication channel and content, price sensitivity and dozens of other dimensions.
Generate, create and use additional data
DnA algos are equipped with creating and dealing with useful additional data: primary marketing research inputs (satisfaction or loyalty, competitive substitutes, SWOT analysis outcomes and priorities, marketing communication awareness, social sentiment, channel importance), social and market related data (demographics, attitudes, interests, lifestyles, etc) to flesh out the marketing analytics results. DnA Data Enrichment processes with DnA Predictive Marketing Engine usually generates other metrics, like peaks and valleys of transactions and units and revenue, seasonality variables, per cent of discounts per customer, top categories purchased per customer, share of categories (per cent of baby products compared to total, per cent of entertainment categories compared to total), calculates time between each purchase, offer and time until purchase, number of units and transactions per customer, time between categories purchased and roughly 200 more attributes.
Apply the algorithm; output profiling – Predictive Marketing Analytics in practice
DnA fast and non-arbitrary algorithms are trying to achieve maximum separation with a view to level a different strategy against each segment. In the profiling, the differentiation of each segment should make itself clear: DnA Analytics demonstrates that the solution does discriminate between segments. The goal is to distinct each segment to create obvious strategy (for each segment). This approach gives you actionable insights fine-tuned to your unique challenges, market position and data.
Scale Model to score database
The Predictive part is to add each customer’s probability to belong to each segment. At the very end of the process each customer will have a probability to belong to each segment and the maximum score wins, i.e., the segment that has the highest probability is the segment to which the customer is assigned. This sounds easy, however DnA Prescriptive Engines spin it and make very sophisticated and unique outcomes, continuously testing selection / targeting, channels, products and messages – and you can track the results and changes if there is any. The output is fully actionable, can be utilized immediately – at any scale and any dimension.
Prescriptive feedback test and learn phase
Setting milestones and experimenting design help to improve your established process, increase your ROI and demonstrate your results to anyone interested (C-level, Shareholders or even Colleagues). Control groups, business as usual milestones will help you to measure your results, tune your system and learn from these processes. Additionally Predictive Analytics in Marketing covers the beginning to experiment with elasticity, price sensitivity, new product / service introduction modelling, combine product categories, introduce new channels and test messaging… and so on… your opportunities now infinite.
Data & Analytics – Predictive Analytics in Marketing 2016-2017.