Personalizing the News: Lessons from 20 years of transforming data into personalized insights for voters, home buyers & Fortune 100 companies

Today, when we discuss personalization most people tend to think of it from a Netflix or Facebook perspective, where the content itself is not personalized, but instead an algorithm recommends content to view.

The much, much bigger opportunity is to provide personalized content that provides insights targeted to our users’ specific situation and needs, e.g. instead of just providing thematic articles about how bad the weather is overall, provide location-specific insights a la the Weather Channel telling them to take cover because a tornado is heading toward their location.

Finally, what our audiences want is not just information, but tools that make it easier to act upon the information provided, either by themselves or in conjunction with their community, e.g. help users to:

  • Find out where their specific elected officials stand on an issue, and enable them to easily give feedback to their aligned politicians.
  • Determine what and when they should plant in their garden, and then streamline the ordering of the seeds and plants online from their local garden store.

Essentially rather than just writing a story about an issue, combining technology and storytelling to help people make better decisions and more easily take action.

Ultimately, in order to achieve our public-service and profit goals, news organizations and journalists will need to embrace the fact that they are actually tech companies in order to deliver the personalized insights and empowerment tools our audiences and communities want.

As a former reporter/editor with 20+ years of experience building decision support tools used by hundreds of thousands of people around the globe to help buy houses, repair cars, manage their supply chains and determine who to vote for, I’ll share some of the lessons I’ve learned as to:

  1. Why personalization is a strategy, not a technology.
  2. The various forms of personalization.
  3. The value of personalization to audiences and organizations.
  4. How to balance system and user control.
  5. Why it’s critical to move beyond personalization to facilitating action
  6. Some tips to help you on your journey to becoming a hybrid content and technology organization.

Now, let’s begin.

Personalization – what is it?

Personalization, rather than being a technology, is a strategy that focuses on delivering tools and content that deliver the most value possible to the individual, that empowers users to live a better life and build a better community.

Personalization isn’t just about deploying some algorithm that learns about your reading habits and suggests additional articles to read. Instead personalization is about building a relationship whereby the news organization continually works to please their users by surprising them with useful and enjoyable gifts.

Ultimately, personalization done right is about putting the user at the center of everything we do, and delivering information tailored to the wants and needs of the user, but done in such a way that helps elevate their community, country and world.

Instead of asking, ”How can we get you to read more articles?”, publishers ask, “How can we make each experience more valuable to you? And what combination of items, content and tools, will enable you to live your best life and enrich your community, country and world?”

Personalization becomes about shifting the focus from pushing out content and products structured around how the news provider is organized and functions. Instead, the focus and question becomes, “How can we help you, our customer, be as lazy as possible and still achieve the maximum success possible?”

Done well, personalization delivers usefulness, simplicity, discovery and control. It brings users insights and tools to help them solve problems and find opportunities effortlessly, often before they know to ask.

Ultimately, personalization is all about balance, providing consistency and diversity, comfort and challenge, usefulness and whimsy. It’s like a friend and mentor, both knowing what you want, and also pushing you out of your comfort zone, challenging you to try new things. It’s a relationship that deepens over time, as users agree to share more about themselves in return for more value and insights.

Personalization – Why is it important?
Personalization done right delivers more value to users, advertisers, our society and news organizations. People are now overwhelmed with content, but often lack time and expertise to find the insights, entertainment and activities they want and need to live the healthy, fulfilling, prosperous lives they desire.

  1. Personalization enables users to cut through the clutter, and find the information relevant to themselves, thereby reducing the time required to search for desired information and experiences.
  2. Personalization drives engagement by serving up interesting content, increasing the amount of time and ideally value a user receives from spending on a site. However, this form of personalization can also be very dangerous to the individual and society, as we’ve seen from the way Facebook and YouTube algorithms work to keep users hooked on their sites, regardless of the impact.
  3. Personalization will be key to competing in the future. Customers expect more and more personalized experiences. The one-size fits all approach to content no longer works, especially as organizations like Nextdoor, Facebook, Google, Craigslist, recipe sites and hundreds of other sites encroach further and further into spaces that used to be largely controlled by news organizations.
  4. Personalization will be key to facilitating new experiences and monetization opportunities. More about this later.

Different flavors of personalization

Personalization comes in many flavors, but I’m going to focus on two broad categories, personalized recommendations and personalized/customized content.

Today, when most journalists talk about personalization, what they really mean is the ability for a site to recommend content to an individual, or what I call the recommendation model of personalization.

The future though is all about creating customized content for the individual, which is truly about moving to a news as software model.

A successful personalization strategy embraces both, but each model requires very different approaches and delivers different outcomes.

Recommendation Model = The core content itself doesn’t change, instead systems analyze the content and the user to recommend content to the user. Netflix doesn’t customize the movies for you, they only customize the recommendations for you.

In this model, minimal changes are made to the way content is gathered and disseminated. An article is still an article written by a journalist, the same way it has for hundreds of years.

The big difference is the implementation of enhanced classification schemas leveraging people and computers so that additional linkages can be found between different pieces of content in order to enable a machine to automatically recommend additional content to a user, based on their behavior, demographics, etc.

Personalized/customized content = The content is designed and delivered in such a way that it’s customized for each group or individual, e.g. the weather on your app is customized based on your location, so that you see the specific temperatures and conditions specific to your area.

Developing personalized content requires a fundamental rethink in regards to how the news is gathered and structured for distribution. Instead of the article being the smallest unit of distributed content, the underlying data that goes into the creation of the story is captured and structured in such a way that it can be used to generate more personalized insights.

Embracing the development of personalized content, requires ultimately embracing news as a software product, where instead of an article or interactive graphic being the end result of all your work, you’re launching a product that can deliver ongoing value, and therefore needs to be developed and managed as a product.

Ultimately, booth types of personalization rely on the implementation of smarter portals/content management systems that have the flexibility to serve up personalized content.

Automation and control

Personalization involves two seemingly contradictory concepts, automation and control. Automation being where the system automatically generates personalized recommendations and content, e.g. Netflix and Facebook automatically display content with seemingly no input from the user.

Personalization is also about control, enabling the user to customize the content they are seeing to meet their desires. A great example of these two concepts are what play out in your weather apps. If you enable the app to identify your location, it will automatically display the weather where you are. At the same time, if you choose you can select different locations.

Spotify and Pandora offer similar capabilities. You can either leave the music selection to the applications, or on Pandora you can choose how the algorithm selects your music, either casting a broad net for songs, or stay within a limited set of music.

Personalization done right, offers both automation and control.

Beyond personalized content, to enabling action

Knowledge isn’t power, action is. Oftentimes, the news leaves people with a sense of despair, or wondering how to move from desire for action to actually taking action. The good news is that once news organizations move beyond focusing on generating content to using information technology to improve lives and communities, huge opportunities arise.

Once you’ve developed personalized content that aligns to your individual user’s specific needs, then the next step is to leverage that information to facilitate communication and actions between that user and different entities, whether they be other individuals, political representatives or businesses.

For example, recipe sites already enable you to order the recipe items required to make a meal, thereby lowering the barriers to the actual making of the meal. If you already know someone’s address and billing information, you can streamline the process even further.

Personalized Content: Getting from here to there

The rest of this post will focus on the key steps required to deliver personalized/customized content for the individual, as opposed to the deployment of recommendation engines.

While recommendation engines require either building or licensing new software and more flexible content management systems, they don’t fundamentally impact the reporting process or how an article/story is created and distributed.

Achieving true personalization is a journey that will require fundamental changes to every aspect of the news business in order to achieve the ultimate benefits. However, like every journey, it starts with one foot forward, and certainly won’t evolve in a straight line.

The opportunities are endless, but success requires integrating personalization into the base fabric of the publishing platform and reporting strategies so that we can combine what we know about our audience with the information/tools we have to deliver the optimum experience.

In the case of political reporting that means building frameworks and supporting data structures that allow you to provide not just high-level thematic stories, but the underlying detail so you can provide personalized insights, e.g. the political representatives along with their associated districts, opinions and actions on key issues.

The second is to build a taxonomy for your data and stories that both machines and people can use to group, search, suggest, etc. The richer the taxonomy with clear, but not necessarily mutually exclusive hierarchies, e.g. personal & cultural Identity might be one large group that consists of subgroups of race, gender, etc.

The third is to capture information about your audience to better understand what challenges, issues and opportunities they care about. Depending on their demographics, e.g. age, gender, wealth, location, as well as the particular issues, politicians, races, etc. you can build a profile about your user.

The combination enables users to more easily find the information they are looking for, and to be able to recommend and personalize content and tools for the individual user.

Below are a few suggestions to start the evolution. The reality is that you will be building an entirely new hybrid content/software startup within your existing organization, so the actual process is a book that has yet to be written.

Step 1: Embrace the idea that news organizations are fundamentally hybrid content and software businesses whose mission is to help people live their best lives and improve their communities. Until news organizations embrace the concept of news as software, legacy constraints imposed by the printing press, TV and radio, will prevent journalists from delivering the value our country and communities desperately need.

Move beyond the story, i.e. article, photo, infographic or video, as being the foundational element of journalism, and instead break the story into constituent components, i.e. figure out the underlying data that goes into the story, as personalization isn’t just about targeting specific stories to users, but personalizing the information that a user receives.

Step 2: Identify a key pain point/or story for your users, especially one that can scale nationally, or even globally. Software tends to have high upfront development costs, and then low ongoing costs, so find a problem where you can acquire the data and build a solution that works both nationally and locally. Also unlike the traditional model of computer assisted reporting, where the goal is to create a one-time story, you’re often building longer-term information products that need to be resourced so they can be maintained over time.

After all it makes no sense to invest in a solution that costs $100K to build, but only works for a single market where you cannot recoup the costs. Instead, identify a national problem, like COVID-19, where the data can be sourced easily enough that you can provide both local and national insights using a single data model and software application.

Step 3: Identify the key data required to help people make better decisions regarding the specific problem you’re trying to help them with.Today, we often capture tidbits of the data to write an article, but don’t capture or store the information for reuse.

For example, with COVID-19, you not only want to know things about the number of infections, hospitalizations and deaths, but information about where people can get tested, wait times for tests, etc. Remember personalization is ultimately all about helping your users solve their problems the way they define them, not necessarily how journalists’ view the issue.

Remember, no data, means no intelligence. Figuring out what data you need and how to properly structure the data so that you can easily update and report against, will typically be your single biggest challenge. DO NOT SKIMP ON THIS SECTION!

Step 4: Determine where and how you’re going to capture the data. In some cases you’ll be able to acquire via an API. In other cases, you may need to build a simple user interface that you can have your reporters, professionals or vetted volunteers enter the data into, etc.

Step 5: Brainstorm and prototype ways to provide the information in a way that both maximizes the probability that users get the insights they want with minimal work, but also gives them the tools to explore the data to find the specific insights they want.

Subtract for success! Develop journey maps outlining your target users’ current experiences regarding how they solve their problems today, and then use that to identify opportunities to eliminate steps and save time, or find areas for people to linger and explore further.

Success will ultimately require fusing the product and news siloes into combined teams working together toward a common vision, and the reality is there is no simple cookie cutter roadmap.

So enough of the high-level mumbo jumbo, give me concrete examples.

The opportunities to provide more personalized insights and tools are endless. All it takes is to put yourself in your users’ shoes and ask, “How can I make this challenge easier and better for them?”

Below are just a few examples, but the opportunities are truly infinite.

The pandemic

Today, there isn’t a single repository where Americans can go to find information about:

  • COVID-19 trends in their community, state, country
  • What rules and regulations are in place
  • Where you can go for testing
  • Polls regarding the local public’s perception of the disease and support/lack of support for various policies to combat the disease
  • Etc.

If you develop such a platform and repository, you not only enable personalization but provide both local, regional and national insights which can be shared and leveraged by local news organizations as well.

Once you have the information structured so that it can be drilled down to the Zip/city/county level and rolled up at the state and country level, you can provide both personalized and aggregated insights for users. Finally, users can subscribe to updates about these kinds of information in their community, or make it easy to search/find the information they need and want based on location and time.

Home & Garden

Different folks live in different types of homes, in different parts of the world, and have different interests, so why not integrate those different data points to push relevant content that varies by location, housing and interests?

It would be very simple to ask if people are interested in gardening, decorating and home maintenance tips, and then use that information in combination with other insights about their homes, geography, etc. to push them personalized recommendations about:

  • What types of plants to plant based on their specific climatic zones, so one reminder to plant lettuce in April for harvesting in June might got to someone in Kentucky, while someone in Minnesota might get a reminder in June to plant lettuce for picking in August. You can also then offer coupons, etc. from home and garden retailers.
  • Suggested home maintenance based on time of year and region, e.g. for folks in the south put mosquito traps in April to avoid mosquitoes in the summer, etc.
  • What types of plants flourish inside for people who live in apartments and condos versus tips on outdoor plants for people living in single family residences in the suburbs.


If you’re like me, you like movies, theater and music, but aren’t passionate enough to keep track of what’s coming out, what bands or shows are appearing locally. Pre-pandemic, every weekend my wife and I would ask each other what we should do, wishing someone could remind us and push recommendations to us to help us plan our weekend.

Instead of forcing me to wade through the many different entertainment options, why not ask me what types of entertainment my wife and I are interested in, and give me the option to sign up for alerts if any particular genres of music or performers are playing nearby, then I’d be eternally grateful. Or as I’m reading an article about a band, film or musical, give me the option to subscribe to alerts if they’re coming to my town, or show me nearby options, that will make it that much easier for me to act upon the reviewer’s recommendations.

And so on

The opportunities to provide personalized solutions are truly endless. All it takes is to put yourself in your users’ shoes, and ask yourself, your audience members and subject matter experts about the challenges they face and the type of information they use to make decisions, and then begin building solutions that lower the barriers to success.

In Summary

Personalization isn’t about just scraping data from users browsing habits and deploying machine learning algorithms to recommend articles. Personalization is fundamentally about cutting through the clutter to provide our users insights, tools and entertainment that help them live their best lives and improve their communities, countries and world.

Success will require rethinking how we practice journalism, restructuring how we deliver the “news”, and ultimately what a 21st-century “newspaper” is. Done well, it will enable news organizations to outflank Google and Facebook, grow revenue, and most importantly empower individuals to improve their lives and build communities.

Kevin Mireles is a bilingual former journalist with 20+ years of product management experience developing analytical systems in the US and around the world that help consumers and businesses make better decisions, faster and easier than ever before.