Event data describes the use of products, websites, software or practically anything else where a user interacts with trackable, measurable or otherwise analysable moving parts.
Tracking this information allows businesses and organisations to delve into the many ways in which users interact with their products and services. It’s the lifeblood of many data strategies and provides motion to otherwise static entity data like users and customers.
Event data must be purposeful – it can’t just be selected at random.
This article examines the purpose of event data and its various uses within a business context.
Table of Contents
What Is Event Data?
In a modern business context, event data is one of two main types of customer data:
- Event data
- Entity data
When you load up various apps from your phone, visit websites or use other software, you’re interacting via a sequence of events by scrolling, clicking, swiping, interacting with certain features, etc. Any interaction or behaviour can create an event that is timestamped and logged with attributes.
For many businesses, event data, also called interaction and behavioural data, is somewhat grouped around various stages of the sales and marketing funnel as well as the user journey, and is used to track everything from awareness and brand discovery to onboarding, engagement and retainment.
Event data allows businesses and organisations to get an idea of whether the user is deriving value from the product, and whether the product works according to their expectations.
Identifying pain points for optimisation, friction, bottlenecks and errors enables data teams to develop solutions. Event data is critical for building better platforms and products, whether that be a website, app, game, software or other product or service.
Event Data vs Entity Data
Event data, also often known as behavioural data when humans are the ones inputting the events, is the foundation of virtually all analytics. Events can be collected and measured even when there are no accounts registered with the product or service.
Outside of a commercial context, event data describes events in everything from climate and medical science to particle physics and chemistry.
Read this guide on what is customer data for a full breakdown of customer entity and event data.
It’s also worth noting that entities do not need to be present for events to be tracked. For example, when someone uses a website, they may not have an account with that business or organisation, but that doesn’t mean that event data cannot be collected on how they utilise the website (e.g. heatmaps, etc).
Conversely, entity data is much more akin to how most people imagine data to be, a large database full of accounts, or users, or other types of entities.
Event data can be attributed to entity data, e.g. different users with different user characteristics may register different events whilst using a product. The under-40s might use different parts of Spotify or YouTube than the over-60s, for example.
So, event data may or may not be connectable to entity data. Event data describes an action or a state and is usually timestamped as events are chronological and happen in sequence (e.g. add to basket, add payment method and buy).
What Format Does Event Data Come In?
Event data often comes in JSON for exchanging data in attribute–value pairs and arrays. For example, JSON is used to send events to mParticle using the HTTP API.
An example of JSON formatted event data might look something like:
{“user_id”: “id132123”, “first_name”: “Bobby”, “timestamp”: 1639486732}
JSON formats might differ depending on whether events are inbound or outbound.
Event data can also be flat or nested.
- If the event data is nested, then a noSQL database might be best
- If the event data is flat, then an SQL database is likely preferable
Where Do You Get Event Data?
In a commercial context, events are often inducted via platforms such as:
- Google Analytics, Mixpanel, Segment, mParticle and Other Platforms: Google Analytics is perhaps the most likely platform that businesses use to measure website event data. Events can be set up to track all sorts of actions aside from the standard link clicks, form submission, etc. So, you could set up events to track inbound calls made via a web link, or video links. You could also measure image clicks or behaviours such as scroll depth. It’s a similar story in Mixpanel and many other platforms, at least in principle.
- Webhooks: Webhooks are useful for asyncronous events using HTTPS to send JSON payloads for tracking. For example, Stripe provide a webhook which sends an app customer data once the customer pays for an item or invoice. This allows businesses to update their own records when someone uses Stripe to pay for their goods or services.
- IoT sensors: IoT naturally lends itself to the creation of event data as the events received by an array of sensors are often chronological and determine the state of a device as it is used. For example, wearables can send event data about someone’s work out, e.g. their fitness level, which allows businesses to target content aimed at their level (UnderArmor did something very similar to this with their fitness trackers).
- APIs: For example, dataforseo.com is an API stack that can provide SERPs and other SEO event data.
Event Data for Marketers: Discovery and Awareness
Think With Google found that some 87% of online purchasing journeys start with an internet search – this is the first event in many digital journeys.
Events become measurable once users interact in some way with your business, brand or organisation, either directly, e.g. via an app, website, social media account or some other product, or indirectly, e.g. when they mention your business on some other channel.
Some typical events marketers can track here include:
- New followers, likes, comments and other social media metrics. These are fundamental in assessing changes to the discovery and awareness of a brand.
- Brand sentiments and mentions. These let brands track brand share of voice and also enables them to respond to either good or bad PR. One excellent example here is Ocean Spray who utilised social media listening to discover a post by a TikTok user called Doggface420, who filmed himself skateboarding whilst swigging from Ocean Spray. The company visited him with a red truck full of Ocean Spray – marketing and PR gold – see video below.
- Website visitors and interactions. Tracking new website users can help evaluate the effectiveness of campaigns, e.g. advertising or content marketing campaigns designed to drive traffic to the main brand website. Measuring website events like click-throughs (CTRs), bounce rates, dwell time, etc, is a whole discipline in its own right. This enables brands, businesses and organisations to tweak and optimise their website for better usability and ultimately, better conversions. Another aspect of this is A/B testing.
Much of this data can be tracked in either Google Analytics or native social media analytics platforms. These are integrable with customer data platforms and customer experience platforms.
Event Data for Product Teams
Event data is collected by product teams and this is where it really comes into its own as a highly chronological, granular and descriptive data medium. Event data here is post-acquisition, so the users have already been acquired, either by making an account, subscribing to something, downloading or otherwise using some software, etc.
In other words, they are no longer faceless users and entity data has now entered the calculations in the form of account or user data, e.g. names, addresses and other personally identifiable information (PII).
Want to improve your data skills?
See the best data engineering & data science books
How to Use Event Data in Activation
Once you have some users to track, it’s time to look at some activation data. Activation occurs when users begin to derive value from your products or services, or otherwise use some of their features. The general goal here is analysing whether or not your users are actually utilising your product in the way you expect or intend. Some common activation events include:
- eCommerce: Products purchased or subscriptions activated
- Taxi: Ride booked
- Music or Video Streaming: Playing music, using playlists, watching shows, etc
- SaaS: Utilising the product’s various components and features
- Online Booking: Booking tickets or events
Activation also helps businesses optimise their sales funnel from discovery through to conversion and activation. It enables a comparison between marketing campaigns, traffic and changes in activation events. E.g. one marketing campaign might result in high traffic to the website whereas another might result in greater conversions and activations.
Below is an example of a targeted email designed to enhance activation and onboarding:
Product analytics tools display activation data as part of the sales funnel, examples include Mixpanel and Amplitude. Mapping the entire user journey is the most complete way to handle product events and behavioural data, but does require a complete modern data stack.
How to Use Event Data in Engagement
Once activated, customers are not guaranteed to remain as customers, not even for longer than a few minutes if they can instantly cancel and take advantage of money-back guarantees. If a user fails to engage with a product after subbing or otherwise purchasing then this isn’t usually a great sign – businesses should look for happy customers that want to use the product.
Event data for engagement spans a few different areas, such as:
Personalised and Targeted Emails
Probably the mainstay of personalised marketing. Personalised emails can trigger upon activity collected by software or apps, or some sort of tracking pixel. A very common example is an automated email that fires after a user has added products to a basket but has not purchased them.
This applies both to activation and engagement. Onboarding emails thank customers for their purchases, sub or sign-up whilst also directing them to the product. This assists the user in extracting value from the product, especially if there are lots of features.
Grammarly’s welcome email below showcases components of the products that users wouldn’t necessarily derive value from if not prompted.
Event data here can combine with entity data to create personalised campaigns targetted at different users, e.g. those who buy regularly and have high customer lifetime value, or those who live in parts of the world who celebrate Christmas, etc. You can drill down into events and entities as much as you like – advanced segmentation is available in products such as Segment.
This allows different types of engagement emails to be sent to different users, e.g. an enterprise customer might benefit from different comms to an individual user.
In-App Messages
In-app messages appear in the software itself. One excellent example here is in-game clues that trigger when events indicate that a player is stuck. The same applies when users use SaaS software, as some users will require assistance as they explore the tool. Here, expected actions events are defined by a user’s flow through the software. If the user deviates from what they’re expected to do, in-app messages can trigger to get them back on the right track.
Push Notifications
Many will be familiar with the kinds of push notifications sent by eCommerce apps, takeaway apps and ride-hailing/taxi apps. These alert the customer of key information even when the app is minimised. Research has found that push notifications steeply reduce cart abandonment and can be targeted to certain customers at certain times of the day, e.g. at dinner time, or near payday.
Push notifications can trigger upon external events – like the time and date – or are activated based on user events, such as adding items to a cart and not checking out.
Ideally, products should exhibit rapid activation and rapid subsequent engagement. This shows that the user is a) quickly utilising a product and its features and b) continuing to use them and remain engaged with them.
How to Use Event Data in Retention
Retention is hugely important. It’s often expensive to acquire customers, especially if the business relies on paid promotions.
Retaining customers ensures that customer acquisition itself doesn’t run at a loss. It also enables businesses to hone in on customers that display the highest brand loyalty and customer lifetime value (CLV). High CLV customers warrant closer attention as retaining them is more profitable.
Churn is also a major problem in subscription box products, but does also apply to most businesses that have a consistent customer flow. Churn describes the number of users that leave the product or service over a certain period of time, usually a month. Most estimates place churn rates at around 5 to 7% on average. Read more about developing data strategies to reduce churn here.
Churn can be analysed to discover what user attributes churners have in common, when they churn and what they do prior to churning, e.g. they don’t use a product in X amount of time.
In contrast to those that leave the business, those who are considered ‘retained’ likely:
- Purchase products each month
- Utilise SaaS software every week or so
- Make purchases through the app every month or less
- Listen to music or videos weekly
Retention does vary, particularly with average revenue per user (ARPU), which is similar to customer lifetime value (CLV). High ARPU businesses need to work harder to retain customers as each customer is higher value than those with low ARPU and more customers.
So, if a SaaS tool costs some $50,000 per year on licence vs just $500 or so, then those high-value customers are likely more expensive to acquire in the first place and more important to retain. This is where combining entity data with event data is critical – you need to know the context in which events operate.
What Are The Disadvantages of Event Data?
Event data is not a free past to omniscience regarding customer and user behaviour. Here are several drawbacks of event data:
- The Data Changes: If event data changes for some reason, e.g. it is delivered in a different format or contains unforseen values and data types, then the code might break, and ultimately, the whole pipeline could fail. In large businesses, it can be difficult to tell when someone has made a change that affects the data pipeline. If a feature is modified or edited which changes event data, but the product team (or other team) is not informed, then the pipeline might fail.
- Raw Data: Event data is usually raw and requires cleaning and possibly even enriching. Since some actions create many events, it can be tricky delineating which ones are actually useful in any way. Event data requires a good deal of manual handling and it’s usually a slow and iterative process to locate the events that actually illuminate user behaviours. This is made simpler by CDPs and advanced platforms, and Google Analytics also does a good job of making simpler events easier to track. Using Python or SQL to clean and aggregate data requires additional engineering time.
- Data Storage: You have to store the data somewhere, depending on the platform used. For example, Google Analytics stores it for you in Big Query, but otherwise, you might have to engineer solutions for storing data in the cloud and making it accessible to multiple tools and platforms.
Summary: What is Event Data, And How Do You Use It?
Event data has many purposes both inside and outside of commercial industries and sectors.
In this case, event data is being discussed primarily in reference to typical business situations where event data is useful for describing how users interact with a brand, business or organisation and their products and services.
To learn how to develop a tracking plan, head here. There are also many more posts on the blog aimed at providing both theoretical and practical guidance to those wishing to enact their own data strategies.
FAQ
What is event data?
Event data literally describes events, which could be anything from someone kicking a ball to someone scrolling on an app. Of course, for it to be event data, it has to be tracked, or measured in some way, thus digitising the information for analysis, interpretation and other uses. Event data is inducted through everything from websites and apps to games and digital interfaces of all kinds. It describes how a user or other entity interacts with that digital system or platform via a series of events or behaviours.
How can I use event data?
Everyone from marketers to data scientists and small business owners can use event data. Firstly, event data allows people to really know how they interact with their website, but it does also apply to social media where event data could be considered as everything from brand mentions to shares, likes, etc. In terms of products or software, event data records what people do when they use that software, e.g. what features they use, when and how they use them, etc.
How do I launch a data strategy?
Define business goals as problems that have solutions, then identify what data can help you find those solutions. Typical problems include boosting awareness and discovery, conversion, increasing retention, etc. Develop a tracking plan to detail what data you’ll need to track, which might be as simple as some metrics from Google Analytics (or even Search Console) or as complex as compiling thousands of data points that describe product usage.
