Why Data is the Oil of the Digital Era

Artificial intelligence is fueled by three major components - data, algorithms, and compute.

In this article, let’s explore data, including questions such as:

• What is data?

• What are some of the applications of data across an organization?

• Who are the end users of your data?

• How do you get started?

The Economist, The New York Times, and Wired all refer to data as the oil of the digital era.

According to AI luminaries such as Andrew Ng, the co-founder of Coursera and Google Brain, “AI is akin to building a rocket ship. You need a huge engine and a lot of fuel. The rocket engine is the learning algorithms but the fuel is the huge amounts of data we can feed to these algorithms.”

In other words, without data, there is no AI. But, data is not only for AI companies. Data is the lifeblood of decision-making in every organization.

What is data?

Most people think of data as the structured tabular data that is organized in a spreadsheet such as Microsoft Excel or a SQL table. Yes, tabular data is a crucial type of data, but there are other types of data that are just as valuable and are worth capturing and analyzing.

For example, consider unstructured text (e.g., emails, Slack conversations, Word documents, and PDFs) or audio (internal meetings and phone calls). There is also image and video data.

With machine learning, the value of this unstructured, non-tabular data could be unlocked.

As you think of your company’s data strategy, you should consider data across all these formats, including how to bring in vendor-managed data into your walls to analyze (e.g., your Salesforce, LinkedIn, HR, legal, finance, accounting, and IT data).

You could also bring in external data to enrich your internal data (e.g., acquiring company or demographic data on your users from an external source).

What are some of the applications of data across an organization?

Marketing

• Track customer engagement behavior including email opens, email clicks, visits to your website and social platforms, time spent on each page, clicks on each page, and the path the user navigates through on your site.

• Analyze which sources lead customers to your website, social platforms, or product page such as the referral site, email campaign, or search engine.

• Understand your customer, including which type of device they use (mobile or desktop), browser type (Safari, Chrome, Edge, Explorer, etc.), location, age, gender, and email address type (personal or corporate, Google or other, etc.).

Sales

• Capture and analyze data on the different personas that may purchase your product, including age, gender, location, education level, income, household status, company they work for, industry they are in, and their professional title.

• Launch campaigns (email, phone, in-person, LinkedIn, etc.) and analyze response rates and follow-through.

• Analyze how engaged clients are with your product over time (clicks, views, buys, etc.) and look for upsell opportunities.

Customer Success

• Analyze customer engagement to preemptively intervene and prevent churn.

• Track customer service response times and customer satisfaction with survey responses.

Product Management

• Track user engagement statistics, most often vs. least often used features, feature requests from users, penetration of the product across various market segments, and user surveys on the product.

• Analyze expected value add of a new feature versus effort required and track progress of delivering on each new feature request.

Data Science

• Analyze the data stored in a unified data model, providing business intelligence dashboards as a byproduct.

• Design machine learning applications based on the data captured.

Engineering

• Track productivity of the engineers, including number and frequency of code commits, bugs, and major software updates.

• Monitor software in production including uptime of applications.

Legal

• Track time to complete legal process from start to finish.

• Monitor status of various documents, including when contracts need to be modified, filed, etc.

Finance

• Analyze trends in payroll, invoicing, and expense management.

• Use data to forecast spend in the future and plan ahead.

HR

• Streamline recruiting.

• Track employee satisfaction.

• Preemptively mitigate turnover.

• Capture performance reviews and feedback.

Without a clear and well-defined data strategy, your company will not get the optimal benefits of the data you have. You may not even capture all the data you truly need to make your business thrive.

Think of all the data points above that matter to each and every single functional area. Without a system to capture all these data points and to store them in a unified data model, this data will (a) not be captured sufficiently well (b) if it captured, it may be siloed and not enable the stakeholders from acting on the data (c) become a competitive disadvantage versus players in your space that have a good data strategy.

Who are the end users of data?

When designing your data strategy, keep your data end-users in mind.

• Will they want written commentary and insight from the data or will they want to work with the raw data directly?

• Are they technical and want the data via API?

• Are they business people and want the data as a dashboard or email report?

You need to keep every layer of your organization in mind, from the individual contributor to the C-suite and the board.

How do you get started?

• Establish the data strategy owner. This is the ultimate responsible party for designing and executing on the data strategy. This individual will seek input from others, but this individual owns the strategy going forward. This includes designing a unified data model for your organization.

• Think carefully about all of your data capture needs. Determine all the data points you would like to capture throughout the organization. Do not leave anything out. You could always trim the list to prioritize effort. Articulate why each data point matters and to whom in the organization.

• Plan how you want to store the data and where (e.g., on premise, in the cloud, etc.).

• Establish best practices for data management and evolution. Designing the right data model is the first step, but adhering to this model over time while evolving it as necessary is even harder.

As a startup, this seems like a daunting task. Not only do you have to build a product and find the right product-market fit, but you have to design your data strategy and develop the data and tech infrastructure that you will need to support your growth.

Fortunately, there is a healthy ecosystem to support the data strategy needs of startups now. And, while data is crucial, it is not a magic wand. Ultimately, you will need a combination of great product + great sales + great data to fuel your organization.

Ankur