In 1999, Amazon pioneered their patented one-click checkout. Estimated to have added billions to their revenue alone, it’s one of their defining features and growth hacks that drive customer engagement.
But most people miss the engine behind of the innovation. Frictionless 1-click checkout helps (and that’s the process that patented), but none of this would work without advanced customer profiling.
Customer profiling builds up an incredibly rich, detailed view of every Amazon customer account.
With customer profiling, Amazon can personalize every message in every channel to deliver the most relevant, most engaging messages. They can scale and automate their customer operation far beyond what other retailers can manage. And this is what drives innovations like 1-click at Amazon which multiply customer engagement and conversions.
Amazon’s customers are made up of data. And so are yours.
It’s not just Amazon who are all-in with customer profiling
Whilst Amazon is an online store, they’re still an online marketing company. With SaaS startups, and account-based marketing, we still have to engage customers (and potential customers) with the same channels and content - webpages, emails, ads. We’re playing the same game.
Where Amazon can differentiate is they can scale, automate, and personalize all their customer communication at every touchpoint over the entire customer lifecycle - from the day you first sign up, you’ve never received an impersonal message. Amazon can manage this at unprecedented scale.
This is only possible by using customer profiling to build a complete picture of every customer, and then using that profile to engage everyone. It’s almost as if Amazon knows each of us as individuals!
And this process, and the strategies derived from it are at the heart of growing SaaS startups too - including those in B2B. By combining customer data across multiple tools owned by multiple teams - sales, marketing, customer success, product - B2B SaaS companies like Mention, Segment, Optimizely, and more are able to scale, automate, and personalize all their customer communication too. As a result, they have breakthrough jumps in engagement with their leads and customers.
In this guide, you’re going to learn how you can do the same with your SaaS startup.
What is customer profiling?
Customer profiling is the process of building a complete picture of every lead, customer, and account - customer centricity! As people fill out forms, their page views and actions get tracked, sales notes get logged against their names, and so on, a complete profile can be built.
This answers two major questions:
- Who are they?
- What have they done?
Customer profiling is for individual people. For instance, when they fill out forms, visit websites, use a product, talk with sales and have notes recorded. Customer profiling for B2B also includes accounts - firmographic data and aggregating actions by people from the company.
The key benefits of customer profiling include:
1. Build a real-time, real-life customer journey map
By gathering the data on each customer, you can join the dots in the customer journey, and understand what their next action ought to be, so you can then send the right message with the right call to action included.
Without customer profiling, the customer journey map becomes separated from data and reality. Unless customer data can tell the story, it isn’t helpful in nurturing the best customer experience - which is the entire purpose of customer journey mapping in the first place.
Think how Amazon can react to your every click, purchase and interaction. This is constructed and computed from real-time customer profile data. With customer profiling, you can implement the same tactics to engage leads and customers too.
2. Do they fit your ideal customer profile?
Your ideal customer profile is the most powerful tool for aligning your entire organization around the right kind of customer.
But without capturing data on each customer (and potential customer) - profiling each customer - the ideal customer profile is hard to largely academic. Your teams can’t react to a specific person since you can’t identify them as a good lead or bad lead.
Amazon probably sells more combinations of product than you, but the principle still stands. “Customers who bought this item” also bought is driven by identifying an ideal customer profile for that product. At Amazon, content and promotion plan follows each customer profile - so should yours.
3. Are they qualified for sales yet?
Lead profiling ultimately results in lead qualification or disqualification. Without qualifying leads pre-sales, sales end up having to manually qualify every damn lead themselves. This chokes up your sales team and collapses sales efficiency.
With a complete profile on each lead, you can deeply qualify them for sales, and hand them over with the full context on why they were qualified. This let’s sales prioritize their time better, and gives them the most relevant conversation opener to create and close sales opportunities easier.
Amazon has a huge B2B arm with their Amazon Web Services unit. Just as with their online store, they build deep profiles on each account. With a largely self-service model, sales doesn’t get involved until account spend and activity warrants it. Amazon saves their sales team from a torrent of small, unqualified accounts, and so should you.
4. How can we segment them to send the most relevant message?
Segmentation works by including data that is on a profile (or not). By profiling each customer, it enables more sophisticated, precise segmentation with more data, and more complete data.
Without customer profiling, tools don’t have the data needed for segmentation. Instead of sending highly targeted messages to defined groups of people, you have to send generic, send-to-all spam which doesn’t perform nearly as well.
Amazon doesn’t email you about every product they stock under the sun - your emails, messages, and customer experience is tailored to your interests, buying habits, frequency of purchase, and so on. It’s personally relevant, so you’re more likely to engage. With customer profiling, you can segment your lists, audiences, and contacts to make your messages as deeply relevant as Amazon can at scale.
5. Automate your sales, marketing, and customer teams.
In the same way customer profile segmentation can decide who gets sent messages, it can also define when these messages get sent and create tasks for your teams (like sales, support and customer success). Deeper customer profiles can automate workflows in different tools.
Amazon automates the vast majority of it’s communication. Outside of big campaigns (Prime Day anyone?), it relies on the customer profile to inform who gets what message and when, and so should you.
6. Scale your customer operations
Profile data also enables you to scale your operation. As your customer base grows, so does your customer database. Customer profiling enables your sales, marketing, and customer-facing teams to work effectively instead of seeing a fragment of the whole picture in their own data silo. The more useful profile data in one place, the easier it is to work effectively.
With a system for profiling and using that profile data for action and messaging (which drives engagement, sales and profit), Amazon can scale their customer operation. Customer profiling is the key to scaling your sales, marketing, and customer-facing teams too.
Customer profiling never stops
Amazon doesn’t bake inefficiency into their customer data management, so neither should you. Customer profiling needs to be an ongoing process that scales with your company. That means this data capture needs to be automated through tools, APIs and databases - and if manual work is needed (like entering sales call notes), this should be made into a process.
Sales, marketing, support, customer success, product - everyone depends on customer profile data to flow.
To do this, you need to create a customer profiling strategy, and assess the different methods of customer profiling.
Customer profiling techniques
Remember, we want to answer two questions: who are they, and what have they done. In our ideal customer profile guide, we talked about the importance of nouns and verbs in your ICP definitions. These can be used to answer these two questions.
Most tools feature a customer profile which answers both of these (including Hull). The “who they are” nouns and details, with a timeline or event stream with their “what have they done” verbs and actions.
Forms submissions are the main and most obvious way of easily collecting data, particularly for the nouns (like names, objects, places, things) in our ideal customer profile. So long as you include an identifier like an email address, we can associate their form submission with their customer profile. You can also capture and include more data here to associate with a person) using hidden fields, and pass data onwards to multiple tools with advanced form handlers (follow our deep dive guide on customer data integration).
Live chat and chat bots can profile customers in a similar way to forms, particularly with multiple choice or closed questions to control the format of the answers given.
Open ended sales questions can get more qualitative answers and a richer context than a form can. This data usually winds up as sales or customer success notes inside your CRM having been classified and sorted.
Customer data mining is the process of finding raw data, like company office locations, timezones, revenues from public databases, and turning that into something more productive.
This can take some technical skill to do at scale to scrape, parse, and load data into the tools you need to use. Which is why you may value using…
Data enrichment providers do the job of data mining, then package and sell that data in a format ready to use. This gives you the benefit of all the rich, publicly available data on the web and elsewhere - so you don’t have to ask leads and customers to fill stupidly long forms which kill conversions and mean you miss sales opportunities.
Web analytics help you answer the questions of what actions people have taken online. With page view data, and any event tracking you have set up, you can profile each visitor, lead, and customer’s actions. Merging this with other data can be hard (see our customer data integration guide)
Email marketing analytics give you actions and engagement with your emails. Email and marketing automation tools typically have solid customer profiling capabilities to track opens, clicks, replies and so on, together with data captured through email signup forms.
Product analytics, like web analytics, can give you customer usage profiling. This requires some tracking setup for your specific product.
Databases for storing and powering your product accounts and features. Since this tends to be owned by your product and engineering team, this is the nearest thing to codifying your business logic - who your users and customers are, how they interact, and how you profile them over time.
None of these methods will be that new - capturing data is not the hard part of customer profiling.
Sharing that data, combining together, then using it as one is an all together bigger challenge.
The reality: Without a solid customer data management strategy for profiling leads and customers, and then getting that profile data into other tools, you’ll be forever constrained by the data you can profile within each tool.
Deep progressive profiling (AKA. Copying Amazon’s customer profiling strategy)
What is progressive profiling anyway?
Progressive profiling is the process of customer profiling over time. Instead of creating many duplicate profiles within your tools each time new data or a new session is added, that data gets combined with an existing profile and other previously captured data.
With all the historical data, you have the full context on every lead and customer ready to act on it. Your messages can be more engaging and personal because you haven’t “forgotten” everything about them.
Progressive profiling is NOT just auto-completing forms.
In B2B marketing, many marketing automation platforms tout the advantages of progressive profiling in forms.
When data is already known, form fields are already filled with that data, or the field is not even displayed in the first place. This shortens forms and can improve conversions. But…
Auto-completing forms barely scratches the surface of what’s possible with progressive profiling.
We want to execute like Amazon does - scalable, automated, personalized communication across every channel over the entire customer journey - which depends on deep progressive profiling.
To see what this can really look like, you have to see what really big tech companies are gathering on you. Amazon makes it a pain-in-the-arse to get a sneak peak at your full profile. But there’s some others you could and should check out.
- Google Takeout - “Your account, your data.”
- Facebook - “Download a copy of your data”
- LinkedIn - “Getting an archive of your data”
Whilst there may be privacy concerns around how much data they gather, it does mean they can deliver you the most personally relevant and engaging content. It’s the same story at Amazon - just in a black box to people outside the company.
As a startup, your different sales, marketing, and customer facing tools will each be profiling customers. Progressive profiling needs to happen across multiple different tools.
Creating a unified customer profile from many tools
Progressive profiling should combine all your customer data into a single customer view. With customer data management, the key is the key.
With combining data, the key is the main piece of data that you can associate other data with. For instance, if you know my domain is hull.io, then you can associate the company name “Hull”, our office locations, and so on.
To combine your data, you need a common identifier or “key” between your tools.
For identifying individuals this may be an email address, user ID, address or something similar. For company accounts, this us usually the domain name, address, or IP address.
So long as you have a common identifier between tools, you can match these up. To start with, using
INDEX MATCH in a spreadsheet and then upload the result, but as we discuss here, the goal is to build a scalable automated process for managing your customer data.
Most customer data integrations work by matching with one identifier - for instance, HubSpot will integrate with other tools and sync profiles based on a common email address.
The problem with this is not all your tools and data will share the same type of data. You’ll have valuable page view data from before the signup - perhaps your latest demo request visited your enterprise pricing page before signing up?
As a result, tools like HubSpot, Salesforce, and Intercom miss out on the ability to combine other kinds of data because their integrations can’t use multiple identifiers. This also results in many duplicate profiles being created from data that isn't merged.
To get to deep progressive profiling across multiple tools, you need to combine different methods of integrating across different keys.
For instance, how do you combine:
- IP address from an anonymous visitor session
- Conversion ID from Google Analytics
- Personal email address from HubSpot
- Company email address from Salesforce
- User ID in your product database
This is hard. But…
With a deeper understanding of your different data types, you can merge all these data types for deep progressing profiling.
For instance, combining:
- Web page views before signup: Pass a Conversion ID in your analytics tool (like Google Analytics) as a hidden field in the form submission for signing up. This associates a session with an email address.
- Associate multiple profiles with the same company: Parse the domain name from a contact, and associate with the domain name of a known account
- Anonymous website visitor to Salesforce account: Reverse IP lookup of the website visitor against a company using the Clearbit Reveal API. This returns a domain name that can be referenced in Salesforce. This powers the Reveal Loop.
Without deep progressive profiling across all your tools, you’ll build data silos. Since tools are owned by teams, you’ll drive a chasm between each of these teams as they work off different data.
Without integrating your customer data, every customer is different to every team.
This means you’ll fail to engage each customer in a meaningful way. Only when you combine your customer data can you scale, automate, and personalize your messages and interactions over the entire customer lifecycle.
This is the secret sauce at Amazon
And increasingly, B2B SaaS startups who use many tools across many teams.
Examples of deep customer profiling at B2B SaaS startups
To understand how B2B SaaS startups are using customer profiling, it helps to look at different customer profiling examples.
Let’s look at three examples:
- Onboarding emails by job title
- Product qualified leads: two keys
- The Reveal Loop
1. Customer profiling for segmented onboarding emails
Onboarding is the crucial stage to activating new users and accounts, but too many startups send the same messages to the same people.
Since different people have different roles within an organization (and you have personas for your most common types of users, right?), segmenting your onboarding emails by job title enables you to target each person based on their role and responsibility with a job-to-be-done message. This is a common quick win for improving onboarding rate.
However, not every startup gathers accurate job title data during signup. Minimizing the length of each form reduces friction and improves conversion rate on visitor-to-signup, but without this data it becomes harder to optimize signup-to-paid.
A common solution to this is to enrich your contacts with job title data from tools like Clearbit.
By appending Clearbit data to every new signup, you can create the segment to deliver more valuable, more engaging content without having to ask (and kill your signup conversions).
This is simple and straightforward to profile because both email tools and Clearbit Enrichment work off the same key - an email address.
Without an integration though, this takes either wiring the APIs together yourself or a manual import and export. See our deep dive guide for all the pros and cons of all seven common methods of integrating data.
The result of this will be the Clearbit Enrichment data inside of your email tool, so you you can build segments and trigger drip email flows by job title.
2. Customer profiling for product qualified leads
Product qualified leads (or PQLs) are where product usage data determines whether a lead is qualified for sales. For instance, if a trial user has hit their trial limits (indicating they’re getting value from the product) then trigger a task for your sales team to reach out.
Since PQLs deliver sales accounts that are already realizing value from the product, converting and closing those new leads and opportunities becomes far easier than starting out cold.
This takes combining product usage data with your CRM. Product usage data usually resides in analytics tools and a product database.
There’s a few challenges to orchestrating the data behind PQLs to work.
The first is relatively straightforward - joining product tracking events to an email address that a user signed up with. Most analytics tools do this by default, and a SQL join will solve the next. If not, then a user ID has to be merged.
The next challenge is to combine the data types. Although the email address is likely to be the common key, the data types that can be associated with each are very different. For CRMs, it’s the nouns describing “who they are”. For analytics and databases, the primary data type is their event stream - the verbs "what have they done". But, CRMs aren’t designed to store hundreds (or thousands) of daily events.
This means the data has to be extracted from the analytics tools or product database, transformed into a property that can be consumed by a CRM (like
active_users) before being loaded into the CRM for Sales to use. Given the scale and speed of the data flow with PQLs, this is best done automatically - see the suggested methods for PQLs in our customer data integration guide.
Finally, associating a user with an account can be hard. In B2B, Users signup and take action, but companies buy things. If your trial users sign up with a personal or freemail account like gmail, you can’t rely on the email domain to reconcile the two.
However, certain data enrichment providers like Clearbit can trace company account data to freemail, so you can load the all-important company account details into your CRM for your sales team to use and qualify with. You could also use the company domain name to prospect for their “real” work email address.
3. Customer profiling for the Reveal Loop
The Reveal Loop is an advanced data flow for account-based marketing. It enables you to identify and target accounts with personalized messages across multiple channels from an anonymous visitor landing on your website.
You can learn more about the Reveal Loop, and how ABM teams are using it to land over $10,000 per day from their account-based programs.
It is customer profiling taken to the extreme with multiple types of keys used together in quick succession - sometimes to return a near real-time result like a live chat message or personalized webpage.
At a high level, here’s how it works:
- Turn anonymous website visitors into a company name.
- Then a company name into a list of prospects from that company.
- And then return all the data about each of these prospects.
To make this work, this uses multiple keys in turn - one key gives the next key gives the next. Like so:
- IP address: Company domain name
- Company domain name: Email addresses
- Email addresses: Enriched email addresses
… and then send those completed profiles to your email, ads, CRM, and other tools to engage the account.
This is advanced, automated, deep progressive profiling in action. Only with an advanced customer data sync engine or engineering setup to connect all these different data types are the results from the reveal loop possible.
How to setup deep progressive profiling at your business
It’s important you start thinking of customer profiling properly from the outset - aspire to be like Amazon!
Build a complete customer profile. Don’t make do with many separate, siloed profiles in different tools. You need to think how your whole set of tools and teams pictures each lead and customer, and how you can build that unified customer profile form all your customer data.
Start by assessing what tools each of your teams are using. Understand what data each tool is storing - what piece of the puzzle does it track and record? What pieces does it miss? Remember the two questions to answer; what are they (nouns) and what have they done (verbs).
With that, identify the keys to data in each tool. Is the common link an email address, or something different. Try to map how each of these tools can integrate together.
With that, you can build a complete, deep customer profile to power all your customer engagement.