Originally published in The Times of India. This article is part 3 of a series called ‘Reality Check on Media Strategy’.

In the fast-paced digital world, running a media company with static business rules is like driving a car while only looking in the rearview mirror. While these rules may have served traditional businesses well, they are ill-suited for the dynamic nature of digital media.

Static business rules, akin to Soviet-style central planning, hinder agility and innovation. In contrast, algorithmic marketplaces operate more like Adam Smith’s “invisible hand,” optimizing information flow and decision-making in real-time.

To further illustrate this, consider another analogy: Running digital media without algorithms is like managing and operating Netflix with the management toolkit that is appropriate for a traditional business, such as AMC Theatres (or PVR Cinema in India).

Here’s why:

Media is a Dynamic Game

Media must hit moving targets.

Consider murmuration, where large flocks of starlings fly in coordinated patterns, creating dynamic and intricate shapes in the sky. Much like the stock market, most aspects of media resemble murmurations — alive and constantly changing. The business of media involves observing reality for what it is and then imitating it.

Here are few examples:

• Editorial teams in newsrooms stay on top of real-world events and report on them. If there is an outburst of violence against Hindus in Bangladesh, a topic they don’t usually cover, then editors must report it.
• Music producers like Rick Rubin remain attuned to cultural movements and trends, often identifying artists and songs with the potential to resonate widely, then creatively collaborating with them to create impactful music.
• SEO is an algorithm that stays on top of what people are searching for on the Internet. Trending on social media platforms is an algorithm that tracks what people are discussing. Distribution teams in media must stay on top of SEO and social algorithms and either produce content that is already trending or make their message trend.
• Ad networks predict the right advertisement to target a specific user at the right price point. If media companies don’t stay on top of changing floor prices, they will lose money.

Media must hit targets moving at different speeds.

For example, news cycles vary by topic, with some steadily rising and then sharply falling, like elections; others, like a major accident, experience rapid spikes and gradual declines. Understanding these nuances is key to effective content distribution.

Media must hit targets moving at different speeds, while balancing trade-offs.

Running a media business is like spinning a top. A top loses its equilibrium unless it is spinning smoothly with the right balance and momentum. Similarly, media operators cannot optimize one thing at the cost of another. They must optimize while consistently balancing trade-offs:

• Content distribution needs to be timely and relevant to current events while still aligning with user interests.
• Search engine optimization (SEO) strategies must evolve with changing algorithms without sacrificing content quality or readability.
• Subscription growth is essential, but not at the expense of average revenue per user (ARPU).
• Striking the right balance between ad density and user engagement is crucial to avoid driving audiences away.

If-Then Business Rules

If-Then Business Rules are standard operating procedures set by managers (humans) based on their intuition, heuristics, manual segmentation, and analytics.

Business rules work well when tasks require a human touch (nursing), are simple (barista recipes), have little optimization gain (cleaning the kitchen), or are best done by humans (news gathering).

They govern processes, decisions, and actions, ensuring alignment with strategy, regulatory requirements, and ethical standards. They bring consistency, efficiency, and compliance across the organization. Business rules offer simplicity, clarity, control, predictability, flexibility, accountability, and easier alignment with ethical standards.

Origin

Between WWI and WWII, the field of Management Science and Operational Research emerged, helping the war effort solve complex logistical and operational problems using mathematical models. After the war, these techniques were applied to business problems to improve decision-making, optimize resource allocation, production scheduling, and other operational aspects of businesses.

The Disconnect From Mathematics

In the 1980s, expert systems rose to prominence, allowing domain experts to encode rigid and deterministic rules as If-Then statements rather than relying solely on mathematical optimization. This shift allowed businesses to automate decision-making based on rules that were more descriptive and less tied to mathematical models. In the 1990s, this field further expanded with the advent of BPM (Business Process Management) software.

Mainstreaming

Coinciding with technological changes, three trends resulted in the mainstreaming of If-Then Business Rules:

• MBAs > Management Sciences: By the 1990s, the Masters in Business Administration (MBA) degree became more mainstream than the Management Science degree. MBAs focused on training generalist managers who could operate these expert systems. This trained an entire generation of managers to think in terms of If-Then business rules.
• The cult of A/B Testing: By the 2000s, the culture of A/B testing took root in Silicon Valley and became mainstream in digital businesses. As being data-driven and experiment-driven became a thing, managers started regularly testing new If-Then business rules to drive growth.
• Analytics & Automation Tools: Sensing the growing demand for If-Then business rules, software companies began shipping analytics and automation tools that facilitated this. For example, Clevertap, a famous marketing automation tool, includes this feature set. Though Clevertap does much more, sometimes managers use it primarily for its If-Then business rules.

Examples of If-Then Business Rules in Media

Most early ad-targeting and content recommendation systems were built with If-Then business rules, for example:

• Manual segmentation e.g., labeling users with a taxonomy
• Heuristics e.g., selling umbrellas in the monsoon
• Contextual targeting e.g., showing utensils next to recipe content

For example, an advertisement for a luxury car might be shown to users aged 35–50 who visited the business and automotive sections of a website in the past month. The assumption here is that someone who reads business news is likely to have the capital to purchase a car, and someone who visits the automotive section has the intent to buy a car.

Similarly, businesses started building content recommendation systems using If-Then business rules, such as using text similarity to show users content related to what they are currently reading.

Analytics is a Backward-Looking Science

Imagine if reality — news cycles, SEO trends, social trends, advertising floor price trends, etc. — is the yellow line. Analytics is always a backward-looking science. It focuses on lagging indicators like page views and sessions instead of leading indicators like upcoming trends. Additionally, analytics is also delayed; it is always a day or week old.

Humans can’t perceive and act on patterns in granular data, so they need to rely on blended data. However, blended data that hides temporal and seasonal nuances results in a one-size-fits-all approach that fails to address the unique nuances of each situation, resulting in missed revenue opportunities.

If-Then Business Rules is a Static Toolkit

Business rules, reliant on historical data and fixed processes, cannot keep up with this dynamic environment. They are akin to bringing a knife to a gunfight; the bullet will hit before one even has the opportunity to swing the knife.

If-Then Business Rules is a Static Toolkit because:

• Bloated Codebase: Business managers will formulate a If-Then business rule and prove growth in a metric through an A/B test that runs for a couple of weeks. Once proven, this If-Then business rule gets committed into the codebase as an established standard operating procedure. However, this assumes that that specific If-Then business rule will continue to sustain that growth into the future.
• Services Business: Overtime, reality changes but the collection of If-Then business rules running the business aren’t updated. The only way to fix this situation is to constantly run regular A/B tests and stay on top of the numbers everyday. However, this converts the business from a product business (earn money while you sleep) into a services business. The team is always beholden to daily or weekly operations. This means the team doesn’t have cognitive and resource bandwidth to work on higher order problems.
• People Dependency: Over time, these If-Then business rules become overwhelmingly large and complex, and institutional memory cannot catch up. This creates an over-reliance on specific individuals for specific tasks, hindering organizational agility.

Why It Matters: Loss of Revenue or Unrealized Gain

The static approach leaves revenue on the table by failing to react and adapt quickly to moving targets, which in turn results in direct loss or unrealized gains. While the direct loss can still be measured, in the absence of Management Science (or mathematical costing models), there is no accounting for the unrealized gains, which therefore go unnoticed.

Let’s look at few examples:

• Ad Density for Each User: Digital media must balance engagement and revenue — too many ads harm the brand, while too few leave money on the table. Balancing ad density dynamically in relation to a user’s journey isn’t a problem that analytics and static If-Then business rules can solve.
• Setting of Floor Prices: Google’s ad platform demonstrates the potential of real-time supply-demand adjustments. It was reported that Google is able to allegedly capture ~50% of the indirect ad topline. In comparison, media companies tend to change floor prices once or twice a week using analytics.
• Different Business Units Cannibalizing Each Other’s Revenue: Without algorithmic distribution, division leaders (for example, Marvel, Pixar, Lucasfilms in Disney) within media companies may negotiate for top-of-funnel visibility (a cost center) instead of optimizing conversion rates (a revenue center), leading to missed revenue opportunities. Media is like real estate, where every pixel should drive engagement or revenue. If a division leader gets their way in the negotiations, they might prioritize their division’s goals, which might generate neither engagement nor revenue.
• Scalability: Managing business rules for each category, user segment, and geography is impractical. Digital media needs scalable solutions that adapt to various user segments and geographies. Algorithms provide the adaptability and scalability necessary to stay ahead in an ever-changing environment.

Hence, Algorithmic Media employs Machine Learning

So what’s the solution? Only machines can match the dynamism of reality or beat machines.

The movie “The Imitation Game” is set in the 1940s during World War II, when the British are trying to break the German encryption called Enigma. Despite dedicating a significant amount of manpower to the task, they couldn’t crack it. Against this backdrop, Benedict Cumberbatch who plays Alan Turing, the father of AI, says in the movie,

“Enigma is an extremely well-designed machine. Our problem is that we are only using men to try to beat it. What if only a machine can defeat another machine?”

Unlike analytics and business rules, machine learning closely follows reality, thereby minimizing loss or unrealized gain. This makes AI suitable for dynamic games like content distribution, ad targeting, and news cycle identification.

Algorithms take in live data feeds and automatically decipher the most appropriate rule to employ at that moment for that user. Once the rule is put into action, it generates a new outcome from which the algorithm gets live feedback.

• Analytics rely on day-old or week-old data, while algorithms rely on live data.
• If-Then business rules are decided by humans and then embedded in code or Standard Operating Procedures (SOPs), continuing to be used even when they no longer benefit the business. In contrast, algorithms dynamically create new rules and discard them when they stop working.

Thus algorithms outperform business rules by enabling:

• Micro-optimizations: Making small, incremental improvements across various aspects of the business.
• Personalization at Scale: Tailoring content, experiences, and advertising to individual users based on their preferences and behavior.
• Data-Driven Insights: Continuously analyzing data to identify trends, optimize performance, and inform decision-making.

Business-wise this results in:

• Optimize Revenue: Capture more revenue through real-time adjustments to pricing, advertising, and content distribution.
• Enhance User Experience: Deliver personalized experiences that increase engagement and retention.
• Stay Ahead of the Curve: Adapt quickly to changes in the digital landscape and stay competitive in the face of disruption.

How important are algorithms for the business of Media?

In 2018, Mark Zuckerberg faced a U.S. Senate hearing focused on Facebook’s handling of user data, privacy concerns, and its role in misinformation and election interference, highlighting the need for greater tech regulation. While I don’t want to trivialize the matter, here’s a joke:

Q: How important are algorithms for algorithmic media businesses?

A: Zuckerberg endured the Senate hearings “For You.”

In this joke, “For You” refers to the personalization or recommender system that powers Facebook’s News Feed.

Conclusion

It’s important to note that this is not an overnight transformation. Building or adopting algorithmic systems requires investment in technology, talent, and a shift in organizational culture. However, the long-term benefits in terms of revenue growth, competitiveness, and adaptability far outweigh the initial costs.

Editorial-led media companies will continue to rely on other algorithmic marketplaces — search engines, social media platforms, etc. — to distribute and monetize their content. However, those that fail to adopt algorithmic systems internally will increasingly find themselves at a competitive disadvantage, struggling to keep pace with the rapid changes in the digital media landscape.

This article was part 3 of a series called ‘Reality Check on Media Strategy’.

—

Want to republish it? This post was released under CC BY-ND — you can republish it as is with the following credit and backlinks: ‘Originally published by Ritvvij Parrikh on The Times of India. The author retains the copyright and any other ancillary rights to the post.