Personalization At Scale: The Definitive Guide

Personalization strategy is the least understood element of the customer journey. Teams understand how a more personal experience will drive more conversions and engagement, but they don't understand how to personalize their entire customer experience at scale - beyond just "Hi {firstname}!"

Orchestration and personalization are two sides of the same coin. Just as a conductor can coordinate dozens of musicians in an orchestra to create an experience greater than the sum of their parts, you can design & deliver 1:1 personalized experiences by orchestrating your tools, teams & data.

To understand the art of how to create personalized experiences at scale you need to understand the science of how it relates to your data. One does not exist without the other.

The mistake marketers make is by jumping straight into tactics. Reviewing long lists of personalization ideas then bolting them onto their customer journey. This creates a fRaNkeNsTaCk of tools, messy & siloed data, and a disjointed customer journey. This is not a strategy.

You need to think holistically about your customer experience - which is why we've only just started discussing how you should deliver your customer experience in the sixth of seven guides. Parts 1-through-5 have been laying the groundwork for what you're about to build.

In the penultimate part of The Complete Guide to Customer Data, we're going to summarize our six-part framework for personalization at scale.

Objectives:

  • Understand how personalized experiences come from tools & data
  • Learn Hull's six-part orchestration & personalization framework
  • Review common examples of orchestrations and apply your own
  • Create a strategy for rolling out orchestrations across your entire customer journey
  • Learn the orchestration & personalization best practices

The Complete Guide to Customer Data

Lesson #1: Personalization is customer experience design

Personalization is not (just) about having ideas.

Personalization is really about designing the unique customer experience everyone who buys & uses your product will have.

Instead of working backwards from lists of personalization ideas from other marketers, and then bolting that into your customers (or potential customers) buying journey, start with the end experience in mind and work forwards to put it together.

"Hi {firstname}" is not personalization, if the rest of message is just generic send-to-all spam.

Cookie-based content recommendations are not personalization, if there is no true context on the rest of the person and where they are in the buying journey.

Abandoned cart emails are not personalization, if you can't define the perfect next or previous step for each individual.

This is why The Complete Guide to Customer Data starts by defining your ideal customer profile & customer journey map. Together, these define the data you need (the "raw" material) that you can shape into a personalized customer experience with your tools & teams.

Instead of getting stuck with trying to bolt personalization ideas together, you need a framework to think about how you create your customer journeys.

Customer experience comes from customer data

Tools create your customer experience (like sending email) from your customer data. Tools use "raw" customer data (attributes & events) to create:

  • Segments (groups/lists/views/audiences) to define WHO the experience is for
  • Templates (dynamic content/merge tags/variables) to define WHAT is messaged
  • Workflows (sequences/cadences/chatbots) to define WHEN and WHERE the experience happens

You need to understand how to use this "raw material" through your tools to deliver experiences. With this, you need to master your tools and channels to craft those perfect moments for your leads & customers.

Think how a chef transforms raw food with kitchen equipment into a finished dish. There's the science of temperature, cooking techniques, timing... but then there's art of creating the 'perfect' dish.

Think how a musician transforms a composition with instruments into music. There's the science of acoustics, but the art of producing the 'perfect' tone, timbre, and feeling.

How a painter transforms paint & canvas into a work of art...

Personalized experience IS art.

Look how it comes together creating 2017's most listened-to album. The moments, the thinking, the decisions...

Notice how the art comes from the science of making sound. The discussion about the "percussive" lyrics, the "color" of different instruments and sounds, and where (and why) the additional touches were added. You see how they join the dots to make a complete song?

Orchestrating your customer journey from joining the dots too. And the joining of dots starts in your mind.

To orchestrate your customer journey, you need a framework to join the dots together across your tools, teams & data.

Lesson #2: Hull's Six-Part Framework for Orchestration & Personalization

You can break down every orchestration & personalized experience into six simple questions.

  1. WHO are you targeting?
  2. WHAT are you saying?
  3. WHEN are you saying it?
  4. WHERE are you saying it?
  5. WHY are you saying it?
  6. HOW will you integrate your tools & data?

Six simple questions will bridge the gap between the customer experience you want to create, and the customer data & tools you have. The dish you want to serve, and the raw food & equipment you have.

With these six questions, you define the rules of your orchestration & customer experience.

To break them down one-by-one...

WHO are you targeting?

The golden rule of personalization; WHO becomes before WHAT. Always start with who your message is for vs. creating a message (an experience) then finding people to match.

(Think of buying a gift for someone. You can personalize it when you start with the recipient in mind, vs. buying a gift, and then finding someone to give it too - that's impersonal.)

Start with the same process to define your ideal customer profile. Define your targeting and tie it to data (attributes & events) which you can use to build segments, lists, audiences, and groups within all your tools - a centralized segmentation tool to manage segments across all your team's tools can help with this.

WHAT are you saying?

With your segments defined, now you need to consider WHAT should be said to that specific group of people.

Though the underlying structure & call-to-action might be similar for many people or companies, the contents of the message might vary dramatically depending on the persona.

To do this, templates with dynamic content to swap out & replace words, phrases, sentences, paragraphs, images, and more. This is only meaningful once you've clearly defined WHO you're targeting.

WHEN are you saying it?

Third, now you know WHO you're talking to and WHAT you're going to say, you need to define WHEN this experience happens - what's the trigger?

In your customer journey map, you'll have key conversion events between stages of the customer lifecycle along each customer journey path. For example, the Trial started event when a new user signs up could trigger an onboarding email sequence.

But conversion events are not the only trigger. Often, the goal of an experience is to proactively nurture someone to take action vs. react to someone who has already taken action.

Most sales and marketing activity is proactive. By using data to define triggers, you can deliver more relevant (and effective) nurturing. For example:

  • Website triggers (like Viewed pricing page)
  • Product triggers (like Trial limits modal viewed)
  • Buying triggers (like G2 Crowd page viewed)

Your triggers (and data layer) need to be as fast as your customer experience works. For example, if you're triggering personalized live chat in the page, you need to be able to sync the data (with all the context to personalize the message) to your chat tool in time to respond. (Learn more about real-time data integration practices like minimizing roundtrip requests).

WHERE are you saying it?

Now you've defined WHO, WHAT & WHEN your customer experience will happen, you need to define WHERE - your channels and tools.

For example, your segment of qualified leads (WHO) you want to send a message with a suggested next action (WHAT) when they login to the product for the 2nd time (WHEN) - via email (WHERE).

To make this possible, you need to be able to sync all the data that captures the context around WHO, WHAT & WHEN into the tool that delivers messages in the channel you've decided. To take the previous example, you need to be able to trigger emails based on a login event with dynamic templates. This is why we prioritize data portability and usability when choosing your marketing tools.

You also need to be able to track your customer's reactions across every channel to update your "single source of truth" to sync across all your tools. This is how you avoid "communication clash" where different tools think your customer is at a different stage of the customer journey.

WHY are you saying it?

Why do you need to create an experience Which orchestrations should you create? Which personalized experiences should you create?

In Part 2 of The Complete Guide to Customer Data, you defined your customer journey map. Not only does this define your high-level event tracking plan and attribution model, but this defines the set of customer experiences you need to orchestrate.

Your customer journey map explains why each experience exists and what the next step is. Orchestrations are created with a goal in mind. For example, your free trial exists to nurture self-service users through to becoming customers.

HOW will you integrate your tools & data?

Finally, you need to understand how to define how to orchestrate the personalized customer experience from your tools, teams & data.

This means understanding the data - attributes & events - that you need to power the segments (WHO), templates (WHAT), and workflows (WHEN).

With this, you'll see whether you can use existing data or whether you need to source, track, or compute data in your "single source of truth" to enable the experience.

You'll also need to define your customer data integration strategy to enable your tools to deliver this experience. How do you connect all your data sources with the tools delivering the message?

With more complex orchestrations, you may need a series of actions to happen with your data like enrichment, transformation, prospection, or even internal notifications. For instance, a new lead might need to be enriched before they enter segments & workflows to deliver a personalized nurture campaign.

Take a break? We'll email you the rest

Lesson #3: Building Orchestrations

The six-part framework gives you a method to describe elements your orchestrations. Now you need to combine them together.

First, you need to layout the start an endpoints. The "if this then that" framework makes this simple and explicit.

  • IF a lead requests a demo, THEN qualify & notify sales
  • IF a trial user invites any users, THEN trigger a follow-up email
  • IF a target account visits your pricing page, THEN trigger & personalize live chat, emails & ads

Note: "If this then that" is helpful for defining orchestrations, but not for implementing. Many tools use workflow logic, but this becomes incredibly complicated to manage anything with combinations of data sources, two-way data flows between tools, and large numbers of tools. Learn more in Part 5: Customer data integration best practices at any scale.

Once you have the "if this then that" statement laid out, then you need to apply the six questions of the orchestration framework. This will add the definition you need to orchestrate your tools, teams, and data. Here are examples for each of the statements above.

Orchestrating demo requests

A simple example is to ping your sales reps when a new demo comes in having run some pre-qualification beforehand.

IF a lead requests a demo, THEN qualify & notify sales.

orchestration-demo-request

Rule Description
WHO Demo requested and Clearbit employee count > 51
WHAT Notification with contact info (from form) & employee count (from Clearbit)
WHEN Send notification as soon as Clearbit enrichment returned after Demo Request
WHERE Send notification on Slack.
WHY To trigger the sales rep to qualify the lead and setup a demo
HOW Capture Demo requested events with Hull.js. Enrich leads with Hull's Clearbit integration. Sync to Slack channel #demo_requests via Hull Slack integration.

Orchestrating Product Qualified Leads

In product-led growth, leads often use your product before talking to sales (for instance in a free trial or free plan). The goal is to trigger sales activity when leads take action that indicates a likelihood to buy (or don't take action and need nurturing).

In this example, let's say inviting users to your product is a strong predictor of likelihood to buy.

IF a trial user invites any users, THEN trigger a follow-up email

orchestration-product-qualified-leads

Rule Description
WHO Trial users after their first User invited event
WHAT Send email to trial user from sales rep with First Name Company Name, link to next step, assigned reps name & calendar meeting link
WHEN Send as soon as first User invited event occurs
WHERE Send email via HubSpot.
WHY To nurture the trial lead to take more actions in the product and buy.
HOW Write User Invited event from backend SQL table with Hull SQL Importer, and convert to Trial Actions attribute using Hull Processor, then sync to a Contact Property in HubSpot with Hull’s integration to trigger a sequence.

Orchestrating Account-Based Marketing

With account-based marketing, you don't always have identified leads (i.e. have their email address) but you can associate anonymous actions with companies (like website page visits, G2 Crowd page views, social activity) using data enrichment services. This can be used to engage leads and prospects from the account.

Notice how you can roll out engagement across multiple channels without introducing significant extra complexity.

IF a target account visits your pricing page, THEN trigger & personalize live chat, emails & ads

orchestration-account-based-marketing

Rule Description
WHO Any Page viewed of pricing page by a Target Account
WHAT Send personalized message to book a meeting
WHEN As soon as target account engages
WHERE Send emails via HubSpot, chat via Drift, and ads via Adroll.
WHY To create a sales opportunity
HOW Capture Page viewed events (inc. URL) with Hull.js, reveal company with Clearbit integration & match with Target Accounts via Salesforce integration. Trigger Drift chat via Web Personalization, HubSpot workflows via integration, & update Adroll audiences via CSV.

Testing Orchestrations

When you build complex data flows for an orchestration, it's best practice to build the simplest end-to-end data flow first. For instance, start with one trigger and make sure that works through to sending the end message.

Unlike software development, marketing tools generally don't have a "sandbox" or "staging" environment which means you're often testing your data flows with live data. Starting with a simplified but still end-to-end, trigger-to-action data flow means you can observe & fix errors without compromising your customer experience.

orchestration-end-to-end-first

End points which trigger internal notifications (like Slack notifications or emails) can be helpful "safe" starting endpoints for designing workflows too, versus triggering automated messaging to leads and customers immediately.

test-flow-then-automate-messaging

Finally, it helps to avoid "blind" fully-automated workflows early on. Test and develop your workflows with Slack notifications, manually triggered campaigns, or manual actions by sales-reps, then move to human reviewed-and-approved semi-automated workflows, instead of jumping immediately to triggering automated action at scale.

manual-then-automate

Again, think of the chefs, musicians, and artists - they taste, listen & look before they ship it out. You're creating an experience. You're creating art.

Lesson #4: Orchestrating your entire customer journey

Now you can craft & test orchestrations, you need to think how to stitch together individual customer experiences into your entire customer journey - everything is orchestration.

In Part 2, you outlined your customer journey map. This divides up lifecycle stages (e.g. lead, opportunity, customer) and customer journey paths (e.g. free trial vs. demo request) to produce a set of unique customer experiences. The goal is to define how you orchestrate the whole system.

Just as with individual orchestrations, it is important to not attempt to do this all at once. We observe most teams start with the marketing-to-sales handoff then move "upstream" of the customer journey. Both the sales and marketing teams should use a common definition of the ideal customer profile to qualify leads.

This has two benefits; it aligns the sales and marketing teams, and builds trust in data-driven action (vs. grunt work, heavy manual research, and unvalidated "rules of thumb").

The types of data flows involved in capturing sources of sales-ready leads, (perhaps enriching and) qualifying them, then syncing them to your sales CRM also lay the groundwork for future marketing-led and product-led sales triggers.

Armed with your customer journey map and the six rules of orchestration, you can begin to expand and orchestrate your entire customer journey.

Think of each rule as a dimension to expand along.

  • WHO: Identify best-fit companies and unique buyer, champion & user personas
  • WHAT: Identify the best messaging to engage with each person at their point in the buying journey
  • WHEN: Identify triggers and key moments in each customer experience
  • WHERE: Maximize engagement by taking the message and experience across multiple channels
  • WHY: Maximize personalization across the entire customer lifecycle by orchestrating each lifecycle stage
  • HOW: Minimize data flows by consolidating your data integration strategy.

Add to orchestrations by expanding along one of these dimensions at a time. Create & optimize messaging for different segments. Replicate an orchestration across another channel. Add an additional data source (like another in-product event) to trigger the same or similar action. Expand your orchestrations incrementally.

Orchestrate all your tools, teams & data with Hull

Hull is a real-time customer data platform. Combine data from all your tools, tracking & databases into a unified customer profile. Then cleanse, enrich, segment, and sync across all your tools.

Explore Hull's customer data platform

Lesson #5: Managing complexity of personalization at scale

As you build out your system of orchestration, you need to be deliberate in how you manage the complexity of all your tools & data.

There are several best practices to manage this complexity:

Create a simple "catch-all" orchestration framework for deciding what data you need, tools you use, and how your teams interact. We advocate an ideal customer profile (clear, common, objective definition to tie your data to) and customer journey map (to define your customer experiences from lifecycle stages and customer journey paths).

DRY. Don't repeat yourself. "Fill in" your customer journey map, and use the six-rules of orchestration. Make sure standardize your naming conventions to prevent duplication of data, segments, workflows, and so on.

KISS. Keep it simple, stupid. Don't tackle everything at once. The fastest teams focus on orchestrating one part of the customer journey at a time, implement a simple first use case, and make sure they test their flows end-to-end so they don't have to review & fix their setup retrospectively.

Define a single source of truth. Make sure you have one place to unify data from disparate sources about the same person & company, cleanse your data, define fallback strategies (to prioritize competing data sources), and then sync across all your tools in real-time.

Build complexity into segments & templates, not workflows. Segments & templates take only one computation and are simpler to build & maintain. Workflows can be great for showing a visual overview of data flows, but they trigger lots of individual computations (which causes latency issues), means your leads & customers have a "state", and become significantly harder to build & maintain at scale.

Hire tools for a job-to-be-done. Instead of trying to orchestrate coordinate wrangle dozens of different (often overlapping) tools, keep a lean stack and fewer moving parts by only "hiring" tools you need for specific, defined roles.

Master your tools. With any art, there's fluency and mastery of the tools needed. So practice like an artist - make sure your team knows how to use your key tools, especially segments, templates & workflows.

Use AI to manage extremes of complexity. At scale, you may end up with many dimensions to your data that become difficult to sift through & define the rules for with human intelligence alone. Artificial intelligence (specifically machine learning) can tackle well-defined problems like lead qualification and web personalization to spot & take action on opportunities you will likely otherwise miss and simplify your set of decisions.

Define a clear customer data integration strategy. As you grow and introduce more tools, teams & data, you need a method of tying data together that doesn't become exponentially complex (like native integrations & workflows pointing between everything) but centralizes your customer data in real-time.

Coming Up In Part 7: Customer data management, operations & governance

So far we've discussed tools and data extensively - what data you need, what tools you need, and how to tie them together.

Even if your tools & data are perfectly structured and organized, you need a strategy for your team.

In the final part of The Complete Guide to Customer Data we'll look at how the best teams manage their customer data, run internal data operations, and create a data governance structure that enables every team to move fast.

Get the next guide in the series

Ed Fry

Prev 'Ed of Growth at Hull, working on all things content, acquisition & conversion. Conference speaker, flight hacker, prev. employee #1 at inbound.org (acq. HubSpot). Now at Behind The Growth

If you've questions or ideas, I'd love to geek out together on Twitter or LinkedIn. 👇