Analysing Customer Journeys to Predict Behaviour Adrian Carr A - PowerPoint PPT Presentation

Analysing Customer Journeys to Predict Behaviour Adrian Carr

A Customer Journey Example All companies try and predict outcomes – e.g. sales or churn etc, or even sub outcome events leading up to the target outcome. These events can be across multiple channels, inbound and outbound, and they can also be trigger events just captured from the data. March January February Outcome Outbound Branch Inbound Account log in Competitor Browsing

It’s complicated Terminology Path = Journey = Sequence All of the above mean ‘what happened between two points in time’ Event = Step = Point All of the above are the ‘what’ in ‘what happened between two points in time’ Even though this is just an example, it is already very complicated. • Just one customer • 17 events • Which events are relevant? • Is the order of events important? • Did some people have the same sequence of events, but not the same outcome? • What about demographics and account data – is that relevant?

Sankey Diagrams Who is / was Sankey? A. A Clever Person in SAS R&D? B. A Russian Mathematician? C. A Railway Engineer? D. A Mild Mannered Janitor?

Sankey diagrams are lovely…..but

…but they aren’t easy to draw in . ppt , and they aren’t simple, and they only ever cover a fraction of the universe so let’s start simply with an arrow to represent time and events on this arrow form the customer journey

This path contains multiple events, across multiple channels (e.g. web, phone, social, etc) Each of the circles represents an event, that could have come from something similar to the diagram below

Some of which are significant objectives or goals of an organisation, which one would want to predict and dis/encourage (e.g. churn, or product sale or conversion, or sub conversion) Examples – • Putting something in basket • Downloading a white paper • Completing purchase • Posting an application form • Calling a call center • Responding to an offer • Accepting an offer • Visiting a store

And some of which are irrelevant when predicting the objective

Dropping the irrelevant events makes a problem simpler

Back to our (now relevant event containing) journey…. cutting the time frame (or length of sequence) of analysis to a more manageable length also makes life more manageable The ‘word on the street’ / ‘grapevine’ is that the length of a journey is best measured in number of events, and is 3- 5 events long.

Focussing on the decision that an organisation can make to influence the objective then becomes an easier task These can be considered as ‘intervention points’ These can be considered to be batch or real time too.

An example Customer responds to offer. Offer of a data snack of 2Gb Customer starts to Customer has less than download a new film 2Gb remaining of their inclusive download that is 3Gb large allowance

…and this then easily extends to multiple intervention points across multiple paths

And sometimes the goal is not achieved, but again, this can form an input to the next decisioning path

…similarly, ‘sub conversions’ can be the objective of an activity, or form the entry to the next path (though of course the customer is just on one journey)

In summary our decisioning is now Our inputs are a distilled set of referenced at the point of paths that are relevant to potential intervention, i.e. the driving a decision that can different times where we can drive a positive outcome take action, with our desire being to influence towards a positive outcome goal And we are driving the goals (or sub goals) that we want a customer journey to lead to

And these customer paths sit as a foundation source of insight into the SAS Customer Decision Hub….which can be optimised

Digging a bit deeper……

So we said before ‘Dropping the irrelevant events makes a problem simpler’, let’s dig deeper into that ‘irrelevant’ definition…

An example of relevant vs irrelevant Positive On first inspection, both events seem predictive – This is some data that records events happening Customer Event 1 Event 2 Outcome there are ten events of each type occurring and (e.g. bill shock, or dropped call), and a positive 1 1 1 1 there are ten outcomes that are also outcomes (e.g. churn, or ‘called call centre’) 2 1 0 1 occurring…..but when you look deeper….. 3 1 0 1 4 0 0 0 5 0 1 0 Question: - which are relevant? 6 0 1 0 7 1 0 1 8 1 1 1 9 0 0 0 10 1 0 1 A. Neither 11 1 0 0 12 0 1 0 B. Both 13 1 0 1 14 0 1 0 15 0 1 0 C. Event 1 Only 16 0 0 0 17 0 1 1 18 1 1 1 D. Event 2 Only 19 1 1 1 20 0 0 0 Total 10 10 10

….looking deeper…. Positive Positive Outcome Hit Rate Positive Outcome Hit Rate Customer Event 1 Event 2 Outcome 0 1 0 1 1 1 1 1 Event 1 Event 2 0 9 1 10% 0 5 5 50% 2 1 0 1 1 1 9 90% 1 5 5 50% 3 1 0 1 Total 10 10 50% Total 10 10 50% 4 0 0 0 5 0 1 0 6 0 1 0 7 1 0 1 When Event 1 occurs (e.g. ‘Bill But when Event 2 occurs (e.g. dropped call), 8 1 1 1 Shock’, the positive outcome the positive outcome only occurs 50% of the time…..the same frequency as when Event 2 9 0 0 0 occurs 90% of the time doesn’t happen. 10 1 0 1 11 1 0 0 12 0 1 0 We can ‘attribute’ the positive outcome occurring to the Event 1 13 1 0 1 14 0 1 0 occurring. We can also say that Event 2 is ‘irrelevant’, and therefore 15 0 1 0 we can ignore it from any path analysis. 16 0 0 0 N.B. In reality, the combination may also need analysing, this is for example 17 0 1 1 18 1 1 1 purposes only 19 1 1 1 20 0 0 0 Total 10 10 10

Many of you will be aware of the attribution techniques / options that exist when considering digital spend….. The successful The lead up events are goals (e.g. known (e.g. customer searched for ‘lovely completed wine’ in Google) purchase) are found One of the traditional methods are used to ‘attribute’ the success to the action None of these are analytical – these are ‘rules based counting, whilst ignoring most of the things that need to be counted’

It is a potentially simple extension to traditional modelling methods cust 1 2 3 4 Goal Cust A A 1 1 1 1 1 1 2 3 4 B 1 1 1 0 1 Cust B C 1 1 1 1 0 1 2 3 Cust C The goal is used as the variable to be predicted, 1 2 3 4 and the events are the The paths can easily be predictive input variables. represented as data, and easily considered in a predictive model This is then a potentially smart way to identify if 4 is Caveat – traditional logistic regression usually only picks out 10-15 variables per goal, truly predictive or not so additional intelligence or other methods should be considered

Why is this different to normal analytics? cust 1 2 3 4 Goal Cust A A 1 1 1 1 1 1 2 3 4 B 1 1 1 0 1 Cust B C 1 1 1 1 0 1 2 3 Cust C 1 2 3 4 The only thing we are missing here compared to path analytics is…. The order of events

Why is this different to normal analytics? cust 1 2 3 4 2,3 3,2 Goal Cust A A 1 1 1 1 1 0 1 1 2 3 4 B 1 1 1 0 1 0 1 Cust B C 1 1 1 1 1 0 0 1 2 3 D 1 1 1 1 0 1 1 Sequence style variables Cust C can easily be created to be 1 2 3 4 represented in a normal model. Cust D One could argue that there is no 1 3 2 4 point in doing path analytics, unless these ‘ordered combination variables’ add more discriminative power over and above existing data

More pictorially…. Only if you build two models – and compare them, will you identify how much the order of the events is actually incrementally predictive.

Credit Card Sales Journeys… ‘New School’ / ‘Digital Marketing’ Cust A Successful 1 3 4 Application Q. Why not simply consider Customer On line browsing model scores as events Applies Email Sent for credit card within the path, i.e. dummy events? Successful Cust B Application 2 3 4 Credit Card Response Score Customer >200 Applies Email Sent ‘Old School’ Marketing

So now our decision hub is driven by both relevant paths and relevant scores Some Paths will be purely event driven Others may benefit from the inclusion of scores (perhaps even make a trigger campaign work better) And other paths are 2 just what we used to call campaigns, based on a score based 2 selection criteria

Questions?