This the final chapter of the series on CDPs (Customer Data Platforms). We will discuss how you might want to evolve your CDP over time. If you missed the previous posts, here's where you can find them.
You shouldn’t try to jump right to the most advanced configuration. There are learnings, validations, and time needed for change management take hold... along the way. Trying to do everything all at once introduces a lot of complexity and assumptions that can sink the deployment.
If your assumptions are wrong, you might have to take things in an entirely different direction
It takes time for people to get used to a new process and see how they want to use new tools/data beyond what you have templated.
Integrating learnings from each level, as new features for the next, ensures you are building tool that address needs and desires of your user base.
This maturity curve has 3 levels now, and as new tech and data comes into market, and processes evolve there will be new levels added.
At each level we:
layer on new sources and types of data
enable marketing with more advanced segmentation and processes
adjust/evolve measurement practices for greater insight
Level 1 - Base
Many companies will start here… Is most of your data is in CRM, MAP, or custom database?
Focus
We start with what should be easiest to get approved and consolidate data we have direct control of... the marketing systems. Traditionally each campaign execution system has it’s own rules and ways to groups companies and contacts for campaigning. With a CDP in place, segments across tools should be unified for consistent personalization and actions
Data
Data types include:
Firmographic
Technographic and Intent which may only be partially updated/enriched
Demographic
MAP response data
Marketing Support
This is a good time to get “your house” in order in terms of data (Firmographic and Demographic tend to be the messy with dupe records and fields and errant values). If you don’t clean it up now (while there is focus on the data) you never will... and the CDP will become yet another data wasteland. Now you have one place to examine data and apply data processing
For Segment membership, product Interest, and persona
Start with Rules initially, as it is fast to setup and lets you test hypotheses that you will validate with Data Science in the next level
List build and personalization should be at a segment level
Similarly, use rules-based lead scoring to prioritize response to send to sale for follow-up
Performance
At this point performance measurement remains simple:
Campaign reporting focuses on improving click through and form submission rates
Attribution is rules based (Linear, U-shaped, W-shaped, time decay, or some combination)
Funnel management revolved around MQL generation (flow) and conversion rate to SAL (quality)
Level 2 - Intermediate
Once we have had Level 1 in place with a good amount of campaign history we can use Predictive Analytics to take over some/all of the rules we put in place. This will validate our good assumptions and correct our bad ones.
Focus
Here we’ll want to broaden our data set with data enrichment for TAM companies so that we can have company level personalization for unknown prospects. We also start integrating some sales data to better inform our analytics, campaigning and automation.
Data
From Level 1, we should have most if not all Marketing data connected. If not, add All response and click-stream data.
If missing, enrich all account in TAM with Firmographic, Technographic and Intent Data
From sales systems get Opportunity (general details) and Account team/Owner info
Marketing Support
In the marketing processes, We enhance Segment membership, Product interest, Persona logic with Predictive Analytics. This is also enhanced with additional marketing response data and Opportunity data like (status, products, etc). Use predictive models to score interest by product, that automation can make use of.
Segments will continue to be the main unit for personalization of content but are enhanced with predictive analytics
both in how the member are populated as well as model assets and their ability to move prospects from stage to stage… Next Best Asset
With all account within the TAM being enriched we can do company level personalization for unknown contacts (where we know the company)
we can use Account Team/Owner mapping for personalization (help warm prospect by introducing their sales team, automate sales process emails...
or Chatbots that pre-qualify and route to appropriate Rep
to prioritize campaign responses for sales follow-up we migrate to predictive lead scoring
Performance
With more data (marketing and some sales systems) pulled into the CDP we can enhance how we look at performance:
The addition of some sales data give us new way to slice marketing reports…
we can see marketing campaign activity coverage by sales team or by sales rep
Campaigns that influence Opptys,
Cost per Oppty
Responses / 1000 touches (Effectiveness)
for Attribution we should move away from arbitrary rules-based model and switch to Machine Learning models for greater accuracy
Funnel reporting adds perspective of sales conversions (SQL and beyond) as a measure of quality of marketing activities
Level 3 - Advanced
Building on Level 2, we add more Sales data, and also add Customer Service data to deepen our understanding, process logic, and analytics capabilities further .
Focus
In Level 3, we start to move from a predictive segment based personalization to truly 1:1 marketing with predictive analytics for content recommendations and analysis
Data
We add any remaining Sales data of interest: contacts/emails/Tasks / Activities / Reponses, Opportunity Contacts, and Relationships between these Contacts
Here we also add interesting Customer Management data: purchases and entitlements, support cases, survey data
Marketing Support
These additional data sources combined with predictive analytics will help us predict Co-workers of Interest (based on their activities and relationship proximities)
We can also personalize content based on 3 levels:
1) this person’s interest/activities
2) people at the same company (or better in same buying group)
3) others with the same buyer persona at similar companies
this determines the next best content that we can present in outbound communications and on the website or by a chatbot
You could take this further and dynamically build a personalized Resource Center that we pre-seed with this Content and content they have viewed and saved… as way to help them as they research.
this can be something they can share with other colleagues
Chatbots can use Sales and Customer Management data to route conversations with context
Performance
With the more detailed Sales stage info and time stamps, we can look at the effects of marketing campaign on Creating Pipeline and Converting pipeline
With the customer Management data, we can calculate lifetime value (LTV) and AVG Annual spend, account growth rate etc
Attribution can be further refined by focusing the modelling data to contacts know to be involved with deals (the Buying group)
The Funnel adds a abstraction layer … not just from from a Lead view (marketing side) but a Buying Group view… giving us a 1:1 comparison of interest generated by marketing and a deal created
currently we can create many MQLs for the same deal…
These are just guidelines to help you find the right level to land on. Everyone's need will be different and in the end it all about addressing the business's pains/challenges/goals. The right mix for you will likely be unique to you.
(Share in the comments challenges you are facing and what you are hoping to put in place)
