Analysis Report Four - Trend Analysis and Value Creation

Executive Summary

Digital platforms have transformed how organizations create and capture value by connecting customers, businesses, and other participants within a shared ecosystem. Marr’s discussion of the seven successful business models of the digital era demonstrates that advertising supported, e-commerce, freemium, marketplace, subscription, aggregator, and crowdfunding models increasingly depend on digital interactions and the data those interactions generate. Platform organizations are well positioned to benefit because they can observe transactions and behaviors occurring across multiple participants in their networks. However, platform scale alone does not guarantee long term success. Zhu and Iansiti identify five characteristics that influence platform sustainability: network effects, network clustering, risk of disintermediation, vulnerability to multi-homing, and network bridging. These characteristics influence both the types of data a platform generates and the opportunities an organization has to create value from that information. Strong network effects, for example, can generate increasing amounts of customer and transaction data as participation grows, while network bridging can allow information gathered in one part of an ecosystem to create value in another.

The Northwind analysis demonstrates these concepts through time trend analysis of product categories and customer markets. Beverages consistently generate the highest revenue among the three categories examined, followed by Confections and Seafood, with relatively stable patterns across the full years of the analysis. Geographic performance shows greater variation. The United States generates the highest revenue during most years, but Germany approaches or exceeds U.S. performance during certain periods, while Brazil also experiences meaningful year to year changes. These findings demonstrate why organizations should analyze platform data across multiple dimensions rather than relying solely on aggregate performance.

Executives should treat platform generated data as a strategic asset. Organizations should continuously analyze activity within their networks, use those insights to strengthen the value provided to participants, and identify opportunities to bridge existing networks into complementary products and services.

Introduction

Digital technology has changed how organizations create value, interact with customers, and generate revenue. Marr identifies seven successful business models of the digital era: advertising supported, e-commerce, freemium, marketplace/platform, subscription, aggregator, and crowdfunding. Although these models generate revenue in different ways, many depend on digital platforms that connect users, businesses, and other participants. These interactions also generate valuable data about customer behavior, transactions, preferences, and engagement that organizations can analyze to improve their business models. The ability to create value from this data is influenced by the characteristics of the platform itself. Zhu and Iansiti identify five characteristics that affect platform sustainability: network effects, network clustering, disintermediation, multi-homing, and network bridging (zhu_iansiti_2019?). Network effects occur when a platform becomes more valuable as additional participants join. For marketplace businesses, this can occur when more sellers attract more buyers and more buyers attract additional sellers. Greater participation also creates more interactions and data that can be analyzed to better understand customers and improve the platform. The value of this data can extend beyond simply tracking transactions. Zhu and Iansiti explain that Amazon uses previous purchasing behavior to improve product recommendations, demonstrating how a platform can learn from the data created by its users (zhu_iansiti_2019?). This is particularly valuable for e-commerce and marketplace business models because information about purchases and preferences can be used to improve recommendations, identify changes in demand, and create a more personalized customer experience. Advertising supported platforms can similarly use behavioral data to better understand audiences, while subscription and freemium businesses can analyze engagement to understand retention and conversion. Other platform characteristics can create challenges that data analytics can help organizations identify. Network clustering occurs when users interact primarily within smaller groups or markets. Zhu and Iansiti use Uber to demonstrate how riders are primarily concerned with the availability of drivers in their own location rather than the size of the platform’s entire network (zhu_iansiti_2019?). Analyzing customer activity by location or segment can therefore help marketplace and platform businesses identify differences that may be hidden within overall performance. Disintermediation creates another challenge when participants connect through a platform but later interact directly, while multi-homing occurs when users participate on competing platforms (zhu_iansiti_2019?). Both can reduce the value a platform captures from its network and make understanding customer and participant behavior increasingly important. Finally, network bridging creates opportunities for organizations to use existing networks and their data to expand into complementary products or services. Zhu and Iansiti describe how Alibaba connected its Alipay payment network with its Taobao and Tmall e-commerce platforms and used transaction information to support additional financial services (zhu_iansiti_2019?). This demonstrates how the data created through one digital business model can support another source of value. Rather than viewing platform data only as a record of past activity, organizations can use it to identify customer needs, improve existing services, and discover opportunities for new products or business models. Together, the readings demonstrate that successful digital business models depend not only on attracting users but also on understanding the interactions occurring within their networks. Platform characteristics determine how participants interact and what types of data those interactions create. Organizations that effectively analyze this data can generate insights that strengthen customer relationships, improve decision making, and reveal new opportunities for value creation.

Outside Research

Platform challenges discussed by Zhu and Iansiti are visible in the current business environment, particularly among large digital companies that depend on marketplace, advertising, and subscription based business models. Apple provides one of the clearest examples because its integrated ecosystem creates strong customer retention and recurring service revenue, but it also creates conflict over how much control the platform should have over developers and transactions.

Mims explains that Apple’s tightly integrated “walled garden” creates high switching costs for customers while also supporting the company’s increasingly important services business (mims_2024?). Apple’s services revenue exceeded $22 billion in the final fiscal quarter of 2023 and represented approximately one quarter of total company revenue (mims_2024?). This demonstrates how Apple combines multiple digital business models, including hardware sales, subscriptions, and a marketplace/platform model through the App Store.

At the same time, Apple’s effors to protect this ecosystem create strategic challenges. In April 2025, the European Commission found that Apple violated the Digital Markets Act’s anti-steering requirements and fined the company €500 million (european_commission_2025?). The Commission concluded that Apple’s App Store rules restricted developers from freely directing users toward alternative offers outside the App Store. This issue closely relates to disintermediation. A platform has an incentive to keep transactions within its ecosystem because allowing buyers and sellers to interact outside the platform can reduce the revenue the platform captures. However, Zhu and Iansiti explain that efforts to prevent disintermediation can make a platform more difficult to use and potentially create opportunities for competitors (zhu_iansiti_2019?).

Apple is also facing broader antitrust scrutiny in the United States. The US Department of Justice and state attorneys general filed a civil lawsuit alleging that Apple monopolized or attempted to monopolize smartphone markets (doj_apple_2024?). The government argued that Apple’s practices can make it more difficult for consumers to switch smartphones and can create additional restrictions for developers and businesses. This connects to the larger issue of platform sustainability. Strong switching costs can help a platform retain customers, but an organization must balance those benefits against the risk that excessive control weakens relationships with other participants in its network.

Similar challenges exist for advertising supported platforms. Marr explains that advertising supported business models rely heavily on user data, including clicks, follows, likes, shares, and other forms of online behavior ((marr_2023?)). In April 2025, a federal court ruled in favor of the U.S. Department of Justice in part of its antitrust case against Google involving digital advertising technologies (doj_google_2025?). The case demonstrates how valuable platform generated data can become within an advertising supported business model. Platforms such as Google connect advertisers, publishers, and consumers while collecting large amounts of behavioral and transaction information that can improve targeting and matching.

Amazon provides another example involving the marketplace and ecommerce business models. Marr identifies Amazon as an organization that combines e-commerce with a marketplace model by selling products directly while also allowing third party sellers to operate on its platform (marr_2023?). The Federal Trade Commission and state attorneys general have alleged that Amazon uses practices that restrict competition among sellers and rival marketplaces (ftc_amazon?). This connects with Zhu and Iansiti’s discussion of multi-homing, in which sellers or customers participate on multiple competing platforms. When sellers can easily use several marketplaces, platforms face greater competitive pressure. However, attempts to reduce multi-homing can also create conflict if participants believe the platform is limiting their ability to compete elsewhere (zhu_iansiti_2019?).

These examples show that the same platform characteristics that help organizations create value can also create strategic problems. Network effects, switching costs, and access to large amounts of customer and transaction data can strengthen marketplace, advertising-supported, subscription, and e-commerce business models. At the same time, excessive control over those networks can lead to regulatory pressure, dissatisfaction among business partners, and greater incentives to use competing platforms.

The more sustainable opportuniy is to use platform generated data to create additional value for participants. Rather than relying only on restrictions to retain customers or business partners, organizations can analyze transaction, behavioral, geographic, and engagement data to improve recommendations, identify changes in demand, personalize services, and discover new business opportunities. Zhu and Iansiti describe this type of opportunity through network bridging, where existing users and interaction data can support expansion into additional markets and services (zhu_iansiti_2019?). In the current business environment, organizations can create stronger platform advantages by using their data to make participation more valuable rather than simply making departure more difficult.

Data Visualizations

1 How does revenue for different product categories change over time?

SELECT CategoryName,
       CAST(SUM(od.UnitPrice * od.Quantity) AS int) AS total_revenue,
       strftime("%Y", OrderDate) AS year
FROM Orders
INNER JOIN "Order Details" od
ON Orders.OrderID = od.OrderID
INNER JOIN Products
ON od.ProductID = Products.ProductID
INNER JOIN Categories
ON Products.CategoryID = Categories.CategoryID
WHERE CategoryName IN ("Beverages", "Confections", "Seafood")
GROUP BY year, CategoryName

ggplot(data = category_trends,
       mapping = aes(x = year,
                     y = total_revenue,
                     color = CategoryName,
                     group = CategoryName)) +
  geom_line()

This visual looks as annual revenue trends for Beverages, Confections, and Seafood. Beverages consistently generates the highest revenue of the three categories, followed by Confections and then Seafood. From 2013 through 2024, revenue remains relatively stable across the categories, although Beverages show greater year to year variation. The categories also follow somewhat similar patterns over time, showing that broader changes in customer activity may influence revenue in addition to demand for individual categories. The lower values at the beginning and end of the analysis should be interpreted cautiously becuase they may represent incomplete years.

This demonstrates how transaction data can create value for e-commerce and marketplace business models. A digital platform could analyze similar product and marketplace business models. A digital platform could analyze similar product level trends to identify changing customer preferences, improve recommendations, determine which products deserve greater visability, or identify categories where additional sellers may be valuable. This relates back to “Why Some Platforms Thrive and Others Don’t”. Amazon, for example, using previous purchasing behavior to improve its recommendations as it gathers more information about customer preferences (zhu_iansiti_2019?). A modern platform could combine transaction data like Northwind’s with searches, clicks, reviews, and other behavioral infromation to generate even stronger insights.

2 How does revenue differ across customer markets over time?

SELECT Customers.Country,
       CAST(SUM(od.UnitPrice * od.Quantity) AS int) AS total_revenue,
       strftime("%m", OrderDate) AS month,
       strftime("%Y", OrderDate) AS year
FROM Orders
INNER JOIN Customers
ON Orders.CustomerID = Customers.CustomerID
INNER JOIN "Order Details" od
ON Orders.OrderID = od.OrderID
WHERE Customers.Country = "Germany"
      AND year = "2020"
GROUP BY month

ggplot(data = germany_month,
       mapping = aes(x = month,
                     y = total_revenue,
                     group = 1)) +
  geom_line() +
  scale_y_continuous(labels = scales::label_dollar()) +
  labs(
    title = "Germany Monthly Revenue in 2020",
    x = "Month",
    y = "Total Revenue"
  )

This looks at annual revenue across three customer markets: USA, Germany, Brazil. The USA generates the highest revenue during most years, but the relative performance of the markets changes over time. Germany approaches or exceeds the USA in certain periods, including around 2020, while Brazil generally generates less revenue but also experiences noticable changes from year to year. Compared with the product category analysis, geographic revenue displays greater variation, demonstrating that overall organizational performance may conceal meaningful differences among individual customer markets.

This relates back to network clustering. Zhu and Iansiti explain that platform networks can consist of relatively independent clusters, such as Uber’s city level networks, where users are primarily concerned with participants in therapy own geographic market (zhu_iansiti_2019?). A marketplace or platform business could use geographic transaction data to monitor the strength of individual customer networks rather than relying only on total platform performance. If revenue or engagement begins declining in one market while the overall platform continues to perform well, analytics could identify the weakness early and allow management to investigate competition, customer preferences, pricing, or product availability.

Recommendations for Industry

Based on the research and data analysis, executives should treat platform generated data as a strategic resource rather than simply a record of past activity.Executives should treat platform generated data as a strategic resource rather than simply a record of past activity. Digital platforms continuously generate information through transactions, searches, purchases, clicks, and other interactions. Organizations should analyze this data over time and across different customer, product, and geographic segments to identify changes that may not be visible in overall performance. The Northwind analysis demonstrates this value by showing relatively stable product category trends while revealing greater variation across geographic customer markets. Executives can use similar analysis to identify where demand is growing or declining and make more informed decisions about marketing, product offerings, and resource allocation. Organizations should also use data to strengthen the value of their networks. Strong network effects can attract additional participants, but growth alone does not guarantee long term platform success (zhu_iansiti_2019?). As participation increases, organizations should use the additional data generated by those interactions to improve recommendations, personalization, matching, and the overall customer experience. This is particularly important for marketplace and e-commerce business models, where understanding changes in customer demand can help platforms connect customers with the right products and sellers. Advertising supported, subscription, and freemium businesses can similarly use behavioral and engagement data to improve targeting, retention, and conversion. Executives should also monitor individual clusters within their platforms instead of relying only on aggregate performance. The geographic differences identified in the Northwind analysis demonstrate how individual customer markets can behave differently even when they are part of the same overall organization. Zhu and Iansiti explain that clustered networks can create opportunities for competitors to establish strong positions within individual markets (zhu_iansiti_2019?). Organizations should therefore regularly analyze performance by geography, customer segment, product category, or other relevant groups. If a particular market begins declining while overall performance remains strong, management can respond earlier by investigating customer preferences, competitive activity, pricing, or product availability. Platforms should also focus on creating reasons for participants to remain within their ecosystems rather than depending primarily on restrictions or high switching costs. Apple’s experience demonstrates that a tightly controlled ecosystem can support valuable marketplace and subscription revenue, but it can also create conflict with developers and regulators (mims_2024?). Disintermediation and multi-homing remain important risks for platform businesses, but organizations can address these challenges by making participation more valuable through better services, lower transaction friction, personalization, and stronger relationships with customers and business partners. Finally, executives should use platform data to identify network bridging opportunities. Information collected through one part of a platform can reveal customer needs that could be served through complementary products or services. Zhu and Iansiti demonstrate this through Alibaba, which connected its e-commerce and payment networks and used transaction data to support additional financial services (zhu_iansiti_2019?). Organizations can apply the same principle by analyzing trends to determine where existing customer relationships could support new offerings or even additional business models. Ultimately, organizations create value from platform data when they turn information into actionable insights. Executives should continuously use tim trend and segmented analysis to understand where their networks are strengthening, where vulnerabilities are developing, and where new opportunities exist. By connecting these insights to the needs of their specific digital business model, organizations can use platform-generated data to improve customer value, strengthen their networks, and support more sustainable growth.

References

(article?){zhu_iansiti_2019, author = {Zhu, Feng and Iansiti, Marco}, title = {Why Some Platforms Thrive…and Others Don’t}, journal = {Harvard Business Review}, year = {2019}, volume = {97}, number = {1}, pages = {118–125} }

(article?){mims_2024, author = {Mims, Christopher}, title = {The Main Driver of Apple’s Success Has Become Its Biggest Liability}, journal = {The Wall Street Journal}, year = {2024}, month = {January}, day = {26} }

(article?){marr_2023, author = {Marr, Bernard}, title = {The 7 Most Successful Business Models of the Digital Era}, journal = {Forbes}, year = {2023}, month = {March}, day = {14} }

(misc?){european_commission_2025, author = {{European Commission}}, title = {Commission Finds Apple and Meta in Breach of the Digital Markets Act}, year = {2025}, month = {April}, day = {23}, url = {https://ec.europa.eu/commission/presscorner/detail/en/ip_25_1085} }

(misc?){doj_apple_2024, author = {{U.S. Department of Justice}}, title = {Justice Department Sues Apple for Monopolizing Smartphone Markets}, year = {2024}, month = {March}, day = {21}, url = {https://www.justice.gov/archives/opa/pr/justice-department-sues-apple-monopolizing-smartphone-markets} }

(misc?){doj_google_2025, author = {{U.S. Department of Justice}}, title = {Department of Justice Prevails in Landmark Antitrust Case Against Google}, year = {2025}, month = {April}, day = {17}, url = {https://www.justice.gov/opa/pr/department-justice-prevails-landmark-antitrust-case-against-google} }

(misc?){ftc_amazon, author = {{Federal Trade Commission}}, title = {Amazon.com, Inc. - Amazon E-Commerce}, url = {https://www.ftc.gov/legal-library/browse/cases-proceedings/1910129-1910130-amazoncom-inc-amazon-ecommerce} }