Understanding Requirements for a Personalization Solution

Fitting the customer experience to the customer — at scale — has been the long-running goal, and challenge, of personalization solutions. 

Lessons from Leaders in Personalization

A recent study, Forrester’s Business Technographics Global Data and Analytics Survey 2018, determined that only 7% of companies have figured out how to compete effectively on experiences. These companies are in hyper-growth mode, and Forrester estimates they will drive $1.8T in revenue by 2021. These are the leaders the rest of us – more than half of us, according to the survey — should emulate.

Our research into achievements in personalization have identified three broad lessons to be learned from today’s leading marketers.

  1. Leaders achieve broad engagement and collaboration among all stakeholders.
  2. Leaders are passionate for insights into customer behaviors and motivations.
  3. Leaders create a culture of experimentation.

We see three categories of requirements that are driven by how these leaders have made strides in personalization.

  • Collaboration
  • Integration
  • AI and Automation

Collaboration Requirements

Because the effort to create, deliver, and manage personalized customer experiences engages people with many roles across many departments in a company, it is important that the personalization solution encourage collaboration. The top three requirements are support for many roles, immediately accessible learning, and enterprise controls.

  1. Support for many roles. User interfaces should be oriented around roles, such as campaign manager, and marketing tasks, such as monitoring and optimizing campaigns. The supported roles should include CMO, customer success, customer loyalty, merchandising, channel marketing, ecommerce, IT, mobile app development, administrator, demand generation, lead generation, and product management, among others.   
  2. Immediately accessible learning. New participants in personalization, or those shifting to new areas, must quickly and independently establish the skills. The solution should offer online learning on how to use the solution to accomplish common tasks. Ideally, it will also offer higher level guidance on how to be effective in personalizing the customer experience.
  3. Enterprise controls. A solution used by a number of people, especially across roles and departments, needs built-in controls to ensure they are not making conflicting changes. Workflow for changes, campaigns, and tests is a big help in this environment.

Integration Requirements

Personalization is not achieved with a tool bolted onto your marketing environment. It is achieved with a broad range of capabilities embedded in most of what marketing does. Integration thus becomes fundamental to the value of any personalization solution. The top three requirements are customer data integration, app integration, and accessible services.

  1. Customer data integration. Personalization requires a broad range of data, including data generated during an interaction,  the customer data that typically resides in several sources within a company, as well as data supplied by third parties. A personalization solution must be able to use all of this data in a systematic and consistent way, add to it, derive insights from it, and share those insights with other customer-touching systems.
  2. App Integration. The personalization solution should have published APIs that suit at least your immediate requirements; server-side and client-side integration; and pre-built connectors for common marketing and sales solutions such as Salesforce and Google Analytics.
  3. Accessible services. Ideally, a personalization solution allows its core services to be used in other apps and channels, via a consistent platform, API, and UI. These core services should allow marketers to request recommendations decisions, request and also share audience definitions, request personalization decisions, use customer attributes, and add to customer attributes.

AI and Automation Requirements

Leaders in delivering personalized customer experiences are developing the capability to personalize any part of any customer interaction. Personalization at scale is a challenge can’t be accomplished at scale unless AI is effective at predicting customer reactions to each step in the experience, and automatically presenting the best next step. The top three requirements are control and automation; shared insights; and real time actions.

  1. Control and Automation. At times you will want to control AI’s decisions and actions, so a personalization solution needs to offer constraints such rules and filters.
  2. Shared insights. ideally, marketers should have insight into what ai has learned about their customers – turning ai internals into recognizable attributes, communicated in human terms. AI decisions should be available to other apps in order to promote more consistent customer experience, either via requests for decisions and predictions, or via data such as audiences or customer profiles.
  3. Real time actions. Automatically delivering the best possible experience within each interaction requires real time decisions, predictions, and actions. AI models should be capable of real time or at least hourly update.

Method: Scenario-based Requirements

Technology is always acquired to improve our processes and results. When we gather requirements for technology solutions, we look at what how we have worked in the past, and what we need now and in the near future to do the same work more effectively.  But technology changes the way we do things, in ways that are not always possible to imagine. As a result, we prioritize requirements that will soon have little value while missing those that will soon seem critical. For example, rather than insist it be easier to cut and paste from this screen to that one, insist that the data transfer be automatic.

A inherent tension in the standard requirements process is simplification vs. context. Managing requirements almost demands reducing each to a bullet or headline entry in a checklist. But getting what you need demands that you retain the context. For example, “Salesforce integration” is useful shorthand, as long as you don’t lose sight of the real requirement, such as “Exchange customer profile data with Salesforce in real time”.

We use a scenario-based approach to requirements because in our experience it is your best hope for success with the requirements process, focusing on what you need in future rather than past aggravations you’ve suffered; and capturing the context — the why, who and when —of the requirement.

Here’s how you use the scenario approach: Your team talks through business activities and goals, with the aim of creating a narrative of ideal scenarios. For a moment, forget the constraints and limitations of today’s tools. What are the most frequent, and what are the most important, outcomes that you pursue? What roles are involved in pursuing those outcomes? In a perfect world, how would your team accomplish those goals? In the ideal, you are not wasting time making up for your tools’s shortcomings, such as trying to reconcile segment definitions or revenue reports, importing or exporting data, inputting results, studying dashboards to identify anomalies and their causes.

Use your scenarios as the basis of requirements, and even more importantly, of vendor demo evaluation. Insist that the demo show us how we would accomplish our goals and do our jobs with the vendor’s tools. Don’t accept a canned demo that is organized around the features of the vendor’s tools.

We recommend starting with your top 3 scenarios. The Table presents a generalized scenario that you can use as a starting point for your own.

Scenario: Director of Marketing is overseeing launch of new product category

Task or Event

What is Success

Requirements to Achieve Success

Test the idea with current customers to find the target audience

Identify target market without negative impact on other results

Testing optimization: MVT testing that also segments audiences, predicts responses, and routes traffic to best performing experience

Identify the attributes of the target segment

Begin to understand how target market compares to other segments

Machine-discovered segment is described with human-meaningful attributes

Machine-discovered segment definition is available/useful to analytics and other marketing tools

Seek more customers to add to segment

Quantify the size of target market

Explore possible similar or related segments

AI/ML analysis that identifies similarities within a group that are predictive of behavior

Based on analysis of customer response and target market, decide to launch the category

Feel comfortable that the range and likelihood of outcomes is clear

AI analysis that quantifies outcomes, probabilities, and confidence intervals

Prepare content for target customer journeys

Customers respond to content by taking the hoped-for actions

Analysis identifies similar segments, which marketing uses to ideate content and paths for target segment

Use the target segment definition in campaigns, both web and email

Meet campaign and launch goals

Testing optimization predicts most effective content and paths for visitors and routes them accordingly

Campaigns can use both user-defined segments and observed segments in matching customers with paths and content

Deploy increasingly personalized campaigns to learn more about target customers and create richer profiles

Positive impact on business results

Improved target segment definitions

Understand how to motivate behaviors of customers

Apply third party and company data in analysis, prediction, traffic routing, segmentation

Real time prediction and analysis

Automated deployment of best experience

Of course, your standard process for requirements gathering can’t be ignored. We recommend the scenario approach as an addition to your requirements process, a way to organize your evaluation, and the best way to focus on what is most important to your future success. 

Published by Sue Aldrich

As a leading authority on worldwide customer requirements, practices, technologies, and governance for personalization, Sue researches the technologies and practices that help marketers get the most useful content in front of customers at the right moment: recommendations, search, discovery, targeted marketing, and web content management. Aldrich is an expert on optimizing the methods that help customers find what they need to make buying decisions and/or to solve problems. She helps clients develop personalization, marketing, discovery, and content management practices that will engage customers and improve results.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: