Customer data governance for scaling, squabbling startups

This is the final chapter of the customer data academy on data governance.

Before we dive into the gnarliest of gnarly topics in data management, first a quick recap of what we’ve covered previously.

In the first chapter, we explained the customer stack model for a simple explanation how your customer data should flow in a closed loop.

In the second chapter, we looked at how to cleanse your customer data to make it immediately useful in all your teams’ tools.

In the third chapter, we discussed how to integrate your tools together to sync your customer data to where it’s needed.

In the fourth chapter, we explored customer data orchestration and how separating your customer intelligence from customer engagement can radically simplify your customer operations.

So far - the guide has focused heavily on data. What data, where it is, and how to use it. But there’s an essential piece we’ve missed so far…

It’s the people, damnit!

Your team is your biggest help and hindrance to your customer data management. Even with the best strategy, tools, and data in the world - your team can botch it all in moments.

It’s worse than herding cats on ecstasy.

Customer data management is the “Wild West”.

Just like the American Frontier in the days of the gold rush, your customer data scattered over a vast landscape of individual tools and teams doing their own thing.

Every customer-facing team and tool needs access to customer data. So every customer-facing tool competes, battles, and negotiates to get the data that they need into the tools they need to use. But with this scattered decision making comes chaos...

No owner? No rules!

Without a ruler, a commander, someone to control what’s going on, then your team is free to do their own thing with your data.

There’s no strategy.

No documentation.

No longevity or plan to what you’re trying to coordinate.

Power vacuum? Yeehaw! Enter the cowboys and cowgirls!

Remember, each team will fight hard to get the data they need in the tools they want to use.

Head ‘em up, move ‘em out. There’s always someone who “knows better”. Someone who has an idea how to pull the data together in a way that makes sense - for them.

“Step aside, son. I’ve setup Salesforce dozens of times before…”

But they’re a cowboy, not a governor.

A cowboy doesn’t assume responsibility for the whole customer-facing team. Sales, marketing, support, success. They just look after themselves and their own use cases.

More cowboys step up and fiddle with the data and data flows. Soon, you’ve dozens of people tweaking and changing your data flows to suit their own needs.

But this town ain’t big enough for the both of us!

It’s inevitable - cowboys start treading on each other’s toes, overwriting, and undoing each other’s work. Enter the showdown and politics of customer data management where one team’s data needs may totally undermine another’s. We call this “data dueling”.

Other cowboys hit the road, heading for pastures new, with their trail of data changes undocumented - a mystery to those left behind.

A power vacuum leads to cowboys.

And cowboys f#ck up your data beyond all recognition. #FUBAR

And drive fear into other cowboys and teams who dare touch their data.

Buckle up, partner. You’re in the Wild West.

Things your customer data cowboys say.

You can tell a cowboy attitude to customer data from a mile off. Less ”yeehaw” and “howdy partner”, more something like this…

“I just dumped it all in Salesforce”

“I think I’ll just test it live in production”

“Can’t get this integration to work?”

“Pfffft. I ain’t reading that documentation!”

“I’m just going to play around with this new tool”

“So what? All I did was change a form…”

“Just smoked all our API credits for the day. Whoops!”

“#FixMyShit”

“I haven't touched the Intercom setup for the record”

“Can I get access to the production database?”

“Why have you f’ing revoked my access to the production database?!”

“Salesforce is THE single source of truth”

“Is this because they aren't technically in Salesforce yet?”

“Just dumped 5000 new prospects in Salesforce, but…”

“OMFG who dumped tons of prospects in Salesforce?!”

“wwwwwtttttfffff”

“this is EXACTLY why I don't want this in Salesforce”

“I don't want to throw a wrench into your plans, but…”

You cannot “manage” customer data.

In the midst of the Wild West, and all this data dueling, you’ve built a Cathedral of customer data plumbing that no one dares to touch for fear of breaking things.

But in the Wild West, there are not enough Sheriffs to go around. You can’t manage this. It’s simply not realistic to dive into the details of each individual conflict, tool, team, and situation.

Following the loudest person (or team) screaming about their “single source of truth” is NOT a strategy.

You need to restore law and order to your customer data management.

You need to govern, not manage.

What is data governance?

Governance sits above management. It forms the leadership and strategy of the big picture.

With that strategy in place, other teams can make decisions and take action, but in a way that’s still helpful to the whole organization.

Data governance enables some control and order without getting stuck in every debate. It’s a scalable, sensible, realistic strategy to customer data management.

Best practices for customer data governance

Data governance is about both the strategy and the leadership.

In the previous four chapters, we’ve introduced the idea of the customer stack - a circular framework for explaining how customer data should flow between sending messages, tracking reactions, recording to profiles, deciding actions, to sending more messages.

In the last chapter on orchestration, we took this a set further by introducing the engagement and intelligence layers. The role of the intelligence layer is to get data into a form that’s immediately useful for engaging with leads, customers, and accounts. (Think lasagna, not spaghetti).

This model gives us is our strategy for data governance and it informs how we should build our teams.

Customer intelligence takes ownership for preparing data

Instead of anyone and everyone being able to change data, we can limit how and what data is prepared centrally. The customer stack model makes this far simpler. It creates one “master” customer profile for each person, made up of profile data from all tools.

Data cleansing and computation is radically simplified if all the computation, cleansing, segmentation, scoring and so on happens in one tool (with the master customer profiles) then is synced out to every other tool.

Data integration is radically simplified if these “master” tools can sync the cleansed and computed data with all the other tools. Not a source of truth, but a system of truth.

Data orchestration is vastly simplified if there’s only one centralized “conductor” directing data between tools.

But this also radically simplifies data governance. Your customer intelligence layer involves one tool that can sync immediately useful data to all your other tools in real-time.

With just one “master” tool, it’s very easy for one team to use, master, and document their processes. With one master tool, it’s suddenly very easy for one team to take ownership of your customer data management.

Engagement teams need to be able to use data, but NOT change data and data flows

If the intelligence layer can prepare and package the data, then the engagement layer doesn’t need to touch it. They can simply plug it into their tools, campaigns, and workflows.

Your engagement team’s tasks become radically simplified:

• Select segments to target emails, ads, or live chat
• Build a highly personalized email, live chat or ad template
• Pull lists of leads or prospects for a campaign
• Build an automated workflow or task trigger in a tool

… which means they can focus their time and effort on engaging with leads and customers, not wrestling with data.

This is the breakthrough!

By separating your data operations from your data users, you pull the responsibility of data management from a vast proportion of people on your team.

In the last chapter we talked about the programming principles of KISS (Keep it simple, stupid!), DRY (Don’t repeat yourself), and spaghetti code. This two-tier team structure helps you apply the same principles to the human side of your customer data management.

KISS instead of exponential complexity. DRY instead of WET (waste everyone’s time). Lasagna instead of spaghetti.

So long, cowboys!

This is the data governance strategy that enables you restore order to your “Wild West” of customer data.

Implementing a customer data governance strategy at your company

First, you need a data owner

If data governance is about leadership, then you lead a leader.

Each team has their data needs. So each team has a data owner who is responsible for getting them the data they need.

The data owners work at the “intelligence layer”. The rest of their team sits at the “engagement layer”, interacting with leads and customers. Data owners feed data and insights to the engagement layer.

Who are your data owners?

There are three ways to identify who the data owner is:

1. You already have a data owner - a Salesforce admin, “Marketing Operations Manager”, or something similar. It’s clear from their title and work that they’re not primarily customer facing (they’re not in an engagement layer) and don’t carry the same responsibilities. For instance, your Salesforce admin probably doesn’t have a quota (or the same quota) themselves as your sales reps.

2. You identify the best-fit data owners - it may not be their formal role, but they’re already responsible for tangential areas like sales enablement or building out campaigns. A sales manager, technical marketer or someone whose spending more time fixing data, and less time focusing on quota, campaigns, or account management.

3. You need to create a data owner - maybe you have a total power vacuum and don’t have a clear data owner to identify and name. You have to appoint them. Just a heads up - this can lead to internal friction and rebellion. Remember, so far it’s been a Wild West. Your cowboys may not see the full picture of data governance - they may just see “bureaucracy”.

How to battle the bureaucracy of customer data governance

Bureaucracy for the sake of bureaucracy isn’t tolerated in most growing startups, but a sensible structure that enables people to do their jobs better is. You want customer data management to feel less like friction, so the data cowboy’s don’t rear their heads once again. You need a working relationship from the get go.

Just like any other area of your organization which matures with hierarchy (design, product, sales, engineering), there are three areas you need to consider:

1. Inbox for capturing and processing requests from your team in the engagement layer. This doesn’t need to be overly complex - whether it’s a project management tool, group inbox, Slack channel or something else. Your engagement layer needs to be clear where they can put requests in that will be read.

2. Projects for sourcing, processing, and testing customer data to meet the inbound request. We discussed this at length in chapter two on data cleansing - preparing data to be immediately useful. This is the responsibility of your data owners in the intelligence layer.

3. Outbox for sending data back into the tools where your teams need it. We discussed this at length in chapter three on customer data integration.. Again, this is the responsibility of data owner in the intelligence layer.

Your data owner needs a data governance tool

Without a dedicated customer data management tool for managing your customer data easily and effectively, your data owner will struggle to prepare the data for your engagement layer. Customer data has many moving parts. You need more than a process. You need a product.

Remember, we want to provide data that’s immediately useful. And quickly.

Anything less than that leaves a power vacuum for your customer data cowboys to step back into.

One data owner? Or many?

It’s unlikely that one person knows all the setup of every core tool you’re using. Instead, you’re more likely to have multiple data owners each responsible for a team or core tool.

For instance, your sales team may focus their toolset around Salesforce. Your Salesforce admin may be your customer data owner. But your customer success and support teams may be working out of Intercom, with a different data owner. Most organizations have multiple customer data owners:

• CRM administrator
• Analytics manager
• Data operations manager
• Support administrator
• Marketing operations manager

They’re in the intelligence layer already, but haven’t yet considered the needs of other teams in the organization and system of data. Customer data sits across the entire customer lifecycle. So, your customer data operation - customer intelligence - needs to become a cross-functional team.

Get started with a "cupcake" project

Your Cathedral of customer data plumbing will NOT collapse in one fell swoop.

Your landscape of cowboys will NOT submit to your superbly simple and sensible strategy overnight.

You won’t be able to access ALL your customer data right away.

It’s right to have bold ambitions to pull all your tools, teams, and data together, but it’s hard to pull off in the real world all at once.

Instead, start with a small first project. Enable an existing team to do something new that they couldn’t before that uses data from many other tools and teams. Use these projects as a chance to build bridges (connecting data together), and add value to existing teams.

• Help sales target better fit leads with product usage data
• Help marketing target leads across every channel by syncing their segments, email lists, and audiences together
• Help sales prospect for relevant leads across an account

---

How to create a customer data governance strategy across your teams

You need to acknowledge there’s a power vacuum amongst your customer data management - cowboys (and cowgirls) with an “anything goes, so long as it suits me!” attitude. This results in chaos, conflict, and a cathedral of customer data plumbing in the long run. You cannot “manage” each individual conflict here in every tool, team, and data point.

Instead, you need to split your teams into customer-facing customer engagement teams and customer intelligence teams. The customer intelligence team prepares the data for the customer engagement team - they take ownership.

You need to identify the customer data owners in your organization. Whose job is it formally or informally (or do you need to appoint someone?) to manage the data needs for their team. Together, this collection of customer data owners forms the customer operations team.

Together, your data owners need a process for handling incoming requests for data, managing it easily, and then syncing this data to the tools where your teams need it. Since there are so many moving parts with customer data management, you need more than just a process. You need a product, like a customer data platform.

Finally, instead of trying to wrestle all your customer data management needs all at once, have your customer operations team work on a small first project and grow their influence and impact incrementally. Create cupcakes, not layers of cake.

---

What is a customer data platform?