Data

Why Progressive Profiling is a sure-fire way to building a better understanding of your customers.

Progressive Profiling is an approach to capturing data that can gradually help develop and build our understanding of a customer and their behaviour over time.
12 minute read
By Nick Jordan

This approach to data capture has a myriad of benefits:

  • Increases conversion rates (CRO)
  • Creates better, more personalised user experiences (CX)
  • Boosts sales (ROI)
  • Accelerates a user's path to purchase
  • Builds better leads and therefore allows for enhanced segmentation
  • Increase brand perception

You can incorporate several data components into an individual person’s profile, including:

  • ‘Transactional’ data, such as purchases made
  • ‘Behavioral’ data, like website pages visited
  • ‘Personal or attributes’ data such as of address or age
  • ‘Identity’ data such as cookies or email address
  • ‘Contextual’ data, such as interest flags

Obviously, the more data you hold on your customer, the greater insight into their preferences and interests you’ll gain, aiding the ability to offer personalized communications, delivering Next Best Action (NBA) with the right message to the right person, in the right channel at the right time.

Customer Progressive Profiling

How do we capture this data?

It’s unrealistic to expect to know everything about your customers from day one, plus people’s circumstances change over time, whether through a purchasing funnel or within customer life cycles. In reality, we actually capture a lot of relevant information as data is produced with every interaction. The majority of the time it’s the case that clients don’t have the foundations in place to interpret the data generated, or the way it is captured isn’t structured in a manner to do so. With this in mind, there are a few areas to consider in aiding profiling:

  • Identity stitching
  • Data Centralisation
  • Taxonomy
  • Filling in the gaps

Identity stitching

Throughout the customer journey there are opportunities to gather information, however with the expanse of channels customers are obtaining their information through, it can be a puzzle to see how these pieces of interactions fit together.

We predominantly see channels operating higher up the funnel, such as media and web working with ‘unknown’ entities, meaning records are just an ID number and we don’t know who they physically are. As customers move further down the funnel, they are more likely to make themselves aware to the brand and becoming a ‘known’ entity, for instance, offering an email address and their name. This provides the additional ability to use channels such as email, SMS or even refined targeting in media.

Customer Journey graphic

However, due to the multitude of differing channels and identities which have been generated for a single customer from their interactions, the information gained can be disparate, inhibiting a full account of their journey, therefore increasing the risk of communications being less relevant or even irritating. For example, if transactional data, such as a purchase, isn’t linked to behavioural channels, such as media and web, the customer will be continually targeted for products they’ve already purchased, wasting budget and harming brand perception that they don’t know their customer. Vice versa, clients might be missing opportunities to upsell, knowing that the customer has been browsing additional products prior to purchase, therefore additional product could be recommended post transaction.

The key is the ability to identify an individual across channels and for the interactions and events they have conducted to be consolidated into a holistic customer journey. With the developments in technical solutions such as the prominence of Customer Data Platforms (CDP) coming to the forefront, identity resolutions are available to join customer records across channels to a singular view via deterministic and probabilistic linkages in real-time. This essentially takes the identifiers or primary keys which are produced within each channel or data set and links them into a consolidated master ID as represented below

Hierarchy graphic

Data Centralisation

Providing the ability to unify the identity of customers across channels and within internal data sets will essentially consolidate and centralise the information, enriching the customer profile and the attributes and traits the customer holds. Additionally, as data is centralised this ensures that the customer profile is updated frequently, accounting for any changes in interests or preferences. This can also simplify client’s technical solutions so that all channel platforms reference one place holding the key information required, rather than hunting around different table and servers, providing the business with one source of truth to make the decision from.

Importantly, now we know who’s who, customer interactions will create an ‘event stream’ of what they have engaged with in a chronological manner across channels. Clients can then use this information to provide insight into customer journeys, forming commonality paths based on similar customer profile traits, aiding the ability for relevant Next Best Action (NBA) which you may have read in this previous post on Salesforce Interaction Studio.

Taxonomy

Now we have a way of knowing who’s who and a place where all customer interactions are stored/updated frequently, it will provide clients the ability to communicate with customers across channels with a continuous story. However, what does it all mean?!

Event stream image

One of the biggest pitfalls we see is the standardisation of taxonomy used across channels e.g. we know the customer has called in and the duration but what was the reason for the call? was the customer happy/angry? what was the outcome from the call?. We see that the customer has opened campaign 'C12DHE33' but now I have to dig out in another system what that relates to and what the customer might potentially be interested in by clicking image 'duwe3n4nudd9'.

As with all data related areas, if you put rubbish in then you’ll get rubbish out. By structuring your data via a taxonomy you’ll be able to ascertain key profile details such as product of interest, product life stage and information gathered so far in their journey. Having this standardisation will distinguish customer traits more easily, plus reduce time and resource in cleaning the data prior to any analysis performed, making the results more robust at less cost.

Rubbish in Rubbish out

Filling in the gaps

With the ingestion of clean and structure interaction data, customer profiles will have a number of attributes populated via indirect methods of data collection. However, there may still be gaps in key data points or poorly populated. This is where direct methods of data collection can be implemented by asking the customer questions. There are some key considerations to think about here such as:

  • What are the key data points we need to populate?
  • How do we go about gathering this information?
  • Upon gaining this information what will it enable us to do, what’s the benefit?
  • Why would a customer provide this information to us, what’s the value exchange?
  • What will change in order to deliver on this value exchange, otherwise customers will lose trust and won’t provide further information

Data collection doesn’t have to be a long boring form of questions which the customer is sent to fill in. Questions can be raised in during BAU communications providing snippets of information which progressively build in depth as customers engage and populate. The use of iconography will also aid the engagement and signals a clear action needs to be taken.

For example, for a logistics company the first question to the customer might be ‘what are you interested in?’ upon selection and providing this information, communications can be tailored to offering more detail in their selection e.g. shipping and cargo. Then the next question to be raised in subsequent communication might be around subject areas within that area of interest, again further refining communications to the customer preference as they populate their selection.

Data Creep

This provide the customer a value exchange that they’re receiving communication which are more relevant to them, plus now that we hold this information communications can also be personalised with imagery and narratives which are contextually relevant too increasing engagement and conversion.

Summary

With a wealth of information available to customers we need to provide information which is relevant and personalised to the customers in order to stand out from the crowd. To achieve this, clients must understand their customers wants and needs to adapt accordingly. Progressive profiling will enable a scalable solution which aids dynamic communications to coincides with changes in customer profile trends. However, clients must lay the foundations in which to gather, interpret and activate on the data which has been generated and provided to them.

Nick Jordan
Senior Data and Martech Strategist, ekino London.
An experienced Marketing Technology and Data strategist with a track record of managing successful projects for numerous multinational companies.

more from ekino

8 minute read
Personalisation at scale to deliver data-driven Next Best Actions (NBA) and Next Best Offers (NBO) with automated customer journeys using Salesforce Interaction Studio.
By Bede Feltham