Moving beyond the dashboard: Hacking your data to fuel hockey stick growth

The success of Growth Hacking has forced a radical re-thinking of our relationship with data

Illustrated rocket shooting up into the sky

Over the last several years, a new breed of data-savvy marketer has emerged from the Silicon Valley start-up scene. Forced to make rapid revenue gains on a small budget to survive, these so-called “Growth Hackers” became experts in interpreting consumer behavior from the digital breadcrumbs left behind by their customers. By learning how to capture and read this data in real-time, these marketer/product manager hybrids were able to make frequent, incremental changes to both marketing and product, driving super-sized growth as a result.

Established companies trying to increase agility, accelerate growth, and expand their digital footprint should take heed. The term Growth Hacking may seem like just another buzzword, but it represents an important shift in the evolution of marketing and product management. This methodology has helped fuel the growth of hundreds of successful start-ups like Airbnb, Uber, Instagram, and LinkedIn, forcing traditional companies to take notice. In this article, we’ll illustrate how re-thinking your relationship with data can serve to unlock growth. We’ll also highlight some of the tools and technologies that can help facilitate this transformation.

Data as a conduit of human interaction

Historically, data technology has focused on tracking large collections of things: people, products, transactions, page views, etc. But in a digital world where every mouse click, keystroke, and finger swipe is a trackable and valuable piece of data, we need to start thinking about data as a stream of human interactions. Tracking a purchase or page view isn’t granular enough to understand consumer behavior. When it comes to achieving growth, nuance matters, and for that we need detailed digital signal that can help us gain empathy at-scale.

Fortunately, modern digital and product analytics tools like Amplitude and Segment were built from the ground up to track human interactions. They were designed to capture a high-volume stream of digital events and make them available in real-time to marketers, product managers, and other downstream systems. In contrast, older analytics tools like Google Analytics (GA) and Adobe Analytics were built to track webpage views in a dashboard. This worked fine when web technology was simple, websites had limited interactivity and the desktop was dominant—but we’re way beyond that today in our mobile-first, post-Web 2.0 world. And while GA and Adobe have technically extended their platforms to incorporate events, they are still webpage-centric at heart, and using them to track to events remains clunky.

Not all data is created equal: customer data fuels growth

If we start to think about data as a stream of human interactions, then we need a better way to capture data about humans. The challenging thing about humans is that we interact with companies over an extended period of time, and through a variety of different channels. Unfortunately, traditional web analytics tools are not well-suited for this type of customer-centered analytics. They are good at measuring anonymous webpage views at a point-in-time, but this tells us very little about the humans we’re curious about. After all, page views don’t buy products or services.

To leverage digital signal to yield growth, we need three things from our data: 1) a way to track interactions from multiple channels, 2) the ability identify the same customer across data sources, and 3) an option for blending anonymous data with known customer data after customers identify themselves. Customer Data Platforms (CDPs) like Segment are purpose-built to do this. Segment ingests customer-oriented event data from dozens of sources (including custom ones) and it creates a unified customer profile that persists over time. (The process of identifying the same user across data sources is known as “identity resolution.”) Modern product analytics tools like Amplitude work well when deployed downstream from a CDP. They ingest events that have been unified through a CDP and provide data analysis tools that allow normal Joes to interrogate customer-oriented metrics like retention, conversion, and cohort behavior. Amplitude also has pre-built integrations for dozens of data sources and it can perform identity resolution on its own, so it can be deployed as a stand-alone platform.

The advantage that purpose-built, customer-oriented data platforms have over general-purpose big data technologies (like data warehouses, data lakes,
and data streaming technologies), is that customer-oriented platforms have inherent domain knowledge about humans. This allows them to provide rich, customer-oriented features out-of-the-box, like identity resolution, privacy safeguards, and customer lifecycle reporting. And increasingly, these tools offer out-of-the-box predictive attributes, such as likelihood to churn, propensity-to-buy, and automated segmentation (called clustering). Developing this type of functionality from scratch using general-purpose data tools would take months if not years.

Amplitude, for example, has a feature called Personas that utilizes machine learning (ML) to automatically cluster customers based on their behavior. This allows product managers and marketers to quickly group users based on how they actually interact with the company, rather than manually creating rules based on how they think users should be segmented. Amplitude has also introduced a robust machine learning architecture under-the-hood called AutoML. AutoML is a full-blown machine learning pipeline that marries Amplitude’s customer-oriented event data with large-scale ML technologies like Apache Spark, Airflow, and TensorFlow. This powerful combination enables auto-generated predictive attributes, thereby transforming growth teams into citizen data scientists.

Re-using analytical data for direct activation

However, to truly harness data for growth, we need to re-evaluate the very reason we use data in the first place. Historically, the goal of analytical data was to populate dashboards for executive decision making. The purpose of this data was to provide management sufficient information to guide changes to the day-to-day operations of the business. Under this paradigm, the dashboard is considered the end of the line for data.

But, the ultimate purpose of data is not to provide management with informative charts (that’s merely an intermediate objective). Rather, the purpose of data is to drive ROI. In the case of the dashboard, ROI is achieved through the completion of initiatives enacted by management, based on their interpretation of the data. The problem is that the time between when data is generated, and the time initiatives are enacted (let’s call this the mean-time-to-activation) can be quite long. First, the data needs to be collected. Then it needs to be transformed and displayed in a dashboard. Next, management needs to interpret the data. Then, an action plan needs to be formulated and agreed upon. Finally, the action plan must be executed. The mean-time-to-activation (MTTA) in this case is typically months or quarters.

Unfortunately, time is a luxury that few start-ups have, so growth hackers developed shortcuts for using data to improve ROI. The first is using analytical data directly for activation. Rather than treating the dashboard as a dead-end, analytical data is repurposed and sent to downstream systems for targeted, high-performing customer communications. Put another way, instead of throwing away all the effort that goes into unifying, cleaning, and augmenting data for dashboards, this data is leveraged to personalize the customer experience across channels. For example, cohorts can be created in Amplitude for analyzing customer behavior like conversion, engagement, and retention. If an analysis identifies a group of customers who are likely churn, the obvious next step is to send them an email or mobile push notification to help keep them in the fold. Whereas Amplitude has pre-built connectors that can trigger this communication in real-time, traditional dashboards can only point you in the general direction—it’s up to you to interpret the data and take action. In the case of Amplitude, MTTA is dramatically reduced because—in addition to providing dashboards—it can directly alter the customer experience through automated integration.

Data as a team sport

The second shortcut growth hackers use to decrease MTTA is liberating analytical data from the grasp of management and putting it in the hands of the front-line workers who directly impact revenue-generating products and campaigns. Rather than waiting for executives to interpret summary data and send down orders from on high, growth-hacking product managers, marketers, designers, and engineers dig into detailed data weekly to validate growth hypotheses, identify friction points, and observe customer behavior. By truly democratizing data in this manner, organizations that embrace growth hacking can radically shorten feedback loops, allowing them to execute significantly more growth improvements than their traditionally-minded competitors.

To do this effectively, though, you need the right tools. If data is locked away in systems that are inaccessible to growth professionals (who are data consumers but not necessarily data experts), then the ability to gain insight and understand customer behavior becomes severely limited. If data experts are the gatekeepers of customer insight, then a company’s ability to learn and adapt quickly is limited by the bandwidth and expertise of a limited few.

Dashboards are a good first step toward exfiltrating data for frontline data consumers (provided that a wide group of users is given access). But dashboards alone don’t allow non-experts to interrogate the data and perform ad hoc analysis, a critical requirement for growth teams rapidly testing hypotheses. SQL is the lingua franca in the traditional data world, but it’s difficult for busy, non-programmer data consumers to learn. Google Analytics and Adobe Analytics have their own data analysis tools, but only specialized digital analysts have the expertise to use them for ad hoc analysis.

Fortunately, there is a better way. Modern analytics platforms like Amplitude are designed from the ground up to be used by non-data-experts. For example, the Amplitude user interface employs a natural language-like, point-and-click interface for building queries, making it intuitive to use from day one. Amplitude avoids the use of obtuse, platform-specific terminology. You won’t see strange terms like eVars, sProps, WHERE clauses, custom dimensions, or goal slot IDs. Amplitude also gives teams the ability to add custom descriptions to things like events (e.g. clicking a button) and properties (e.g. the color and text of the button). This gives teams the ability to create a cheat sheet that’s baked right into the tool itself. Amplitude also provides a wide range of pre-made, high-value, easy-to-configure charts, making it easy to start exploring customer behavior out of the gate. Finally, Amplitude provides a host of team collaboration features, allowing growth teams to add comments to charts, publish their analysis, and discuss data within the tool itself.


Although there have been substantial improvements in how we store and retrieve data over the last few decades, the way we consume data has changed very little. The success of Growth Hacking is forcing us to radically reconsider how we use and think about data. Taking a page from the Growth Hacking playbook, product managers, marketers, and data analysts charged with driving growth should focus on implementing four key changes:

  1. Harness digital signal from customer interactions to gain empathy at-scale.
  2. Leverage purpose-built tools that specialize in customer data.
  3. Re-purpose analytical data for direct activation.
  4. Democratize data and ensure that it’s easy to visualize and interrogate.

Does this mean we should discard our existing data analytics tools like Google Analytics, Adobe Analytics, data warehouses, data lakes, and traditional dashboards? Absolutely not. Each tool serves a specific purpose. However, when it comes to driving growth through rapid experimentation and optimization, a more modern toolset and agile methodology are required for achieving results.

Related Articles: