Platform Revolution

Leaders have always needed to focus on a handful of key metrics to guide them. This has been true for thousands of years in every domain of human activity, from business to government to warfare. Consider Jonathan Roth’s description of the key factors for Julius Caesar’s army in the Gallic campaigns (58–50 BCE):

The Roman army took a vast array of materiel into the field: clothing, armor, edged weapons, missiles, tents, portable fortifications, cooking gear, medical supplies, writing materials, and much more … Yet, approximately ninety percent of the weight of the supplies needed by an ancient army was made up of only three elements: food, fodder and firewood. All military decisions from the basic strategic concept to the smallest tactical movements were affected by, and often determined by, the need to provide these supplies to the army.¹

Given a certain number of soldiers and animals, Caesar’s quartermaster could quickly determine how far the army could march and how long it could campaign before reprovisioning, simply by tabulating the quantity of food for the men, fodder for the animals, and firewood for warmth and cooking. These three key metrics shaped many of Caesar’s most fundamental strategic choices.

Leaders of traditional for-profit companies that have linear value chains (pipelines) have, similarly, achieved success by working with a relatively limited set of standard metrics. For example, firms that manufacture goods such as automobiles or washing machines must source raw materials or subassemblies and then assemble these into complete products that are offered for sale to end customers through a variety of sales and marketing channels. The details of the work may be very complex, but as long as the revenue exceeds the total cost of compensating pipeline participants, along with a margin to justify the risk and cover future development costs, all is well. While line workers and middle managers all along the pipeline need to be immersed in the fine points of design, manufacturing, production, marketing, and delivery, leaders at the C-suite level, as well as board members and outside investors, can focus on a few key numbers to get a quick sense of the relative health of the business.

Traditional measures from pipeline businesses, which are familiar to most managers, include cash flow, inventory turns, and operating income. These work in combination to create a useful, broad-brush picture of a business, and their simplicity and clarity helps company leaders stay focused on the factors that are crucial to long-term success rather than getting distracted by secondary details.

Unfortunately, the traditional metrics used in organizing and running pipeline businesses quickly break down in the context of a platform—and developing alternative metrics that effectively measure the true health and growth prospects of a platform business is far from easy.

Consider the story of BranchOut. Launched in July 2010, BranchOut was a professional networking platform based mainly on an app that enabled users to make job-hunting connections via Facebook. Think of BranchOut as a variant of LinkedIn piggybacked on the vast Facebook network. In a world where the majority of jobs are filled not through help-wanted ads or Internet postings but through friend-to-friend word of mouth, BranchOut struck many people as a brilliant innovation. Its founder and CEO Rick Marini managed to attract $49 million in three rounds of investment.

The company’s rocket-propelled ride to the top of the professional networking world was astonishing. BranchOut’s user base expanded from fewer than a million to a whopping 33 million between the spring and summer of 2012. But its implosion occurred as quickly. Before another four months had passed, the number of members had plummeted to fewer than two million. And by the following summer, the company was groping for an entirely new business strategy, hoping to become a “workplace chat” platform that teams of coworkers could use to stay in touch. Rich Marini admitted to reporters that “active users aren’t very big right now,” while saying that BranchOut was “not a failure, it’s still alive.”²

Postmortems pointed to a variety of reasons for BranchOut’s collapse. Some blamed changes in Facebook’s app developer platform, which put a crimp in BranchOut’s communications system. Others pointed out that the whole notion of blending job search capabilities with Facebook’s social networking ambience was misguided. “Looking for a job is stressful,” as one observer put it. “It’s a lot of work. When I hang out with my friends, the last thing I want to do is talk about my job search. I want to escape it.”³

These may have been factors in BranchOut’s failure. But the most significant mistake BranchOut made appears to have been focusing on—and measuring—the wrong things. Flush with investment capital and riding an incredible upsurge in “active user” enrollments, during those fateful months in the middle of 2012, BranchOut kept directing its efforts to boosting membership numbers. It incentivized users to invite as many friends as possible, and made it easy for Facebook members to invite everyone in their network to join BranchOut. As hundreds of millions of invitations flooded cyberspace, BranchOut’s enrollment figures skyrocketed.⁴

Having a person’s name and email address on a membership list doesn’t promise success for a platform. What matters is activity—the number of satisfying interactions that platform users experience. If BranchOut had tracked the activity numbers as diligently as it tracked membership, it might have realized that its millions of members weren’t finding much value in the service—which led, of course, to plummeting membership rolls.

The BranchOut story illustrates a vital truth about the world of platforms. Just as platforms transform traditional value chains, competitive strategy, and management techniques, they also demand new forms of internal measurement.

Let’s return for a moment to the metrics that are most commonly used by pipeline managers: key numbers like cash flow, inventory turns, and operating income, as well as ancillary metrics like gross margin, overhead, and return on investment. These tools, in their varied ways, help to measure the same thing: the efficiency with which value flows through the pipeline. A successful pipeline business is one that produces goods and services with minimal waste of resources and then delivers a large quantity of these goods and services to customers through well-managed marketing, sales, and distribution systems, thereby generating revenues that are more than sufficient to recoup costs and produce profits to reward investors and finance future growth.

Pipeline metrics are designed to gauge the efficiency of this value flow from one end of the pipeline to the other. They help managers recognize bottlenecks, logjams, and breakdowns in the flow that require improved process efficiencies or system enhancements that will facilitate a larger, faster, and more rewarding stream of value through the pipeline. Thus, when a statistic like inventory turns unexpectedly plummets, it’s generally a sign of overstocking, product obsolescence, or marketing failure, while an excessively high rate of turnover may indicate understocking and consequent loss of sales. Carefully monitoring this metric can help managers make necessary adjustments to keep the business humming.

This kind of (admittedly simplified) analysis doesn’t work when we shift our focus to a platform business. As we’ve seen, platforms create value primarily through the impact of network effects. Platform managers in search of metrics that reveal the true health of their business need to focus on positive network effects and on the platform activities that drive them.

In specific, platform metrics need to measure the rate of interaction success and the factors that contribute to it. Platforms exist to facilitate positive interactions among users—particularly between producers and consumers of value. The greater the number of positive interactions the platform creates, the more users will be drawn to the platform, and the more eager they will be to engage in activities and interactions of various kinds on the platform. Thus, the most important metrics are those that quantify the success of the platform in fostering sustainable repetition of desirable interactions. The end result: positive network effects and the creation of enormous value for everyone involved, including the users of the platform as well as the sponsors and managers of the platform.

Note the difference between this core metric and the core metric of the pipeline. Whereas a pipeline manager is concerned with the flow of value from one end of the pipeline to the other, the platform manager is concerned with the creation, sharing, and delivery of value throughout the ecosystem—some occurring on the platform, some elsewhere. For a platform manager, process efficiencies and system enhancements may be quite important—but only insofar as they facilitate successful interactions among users. The big goal on which platform managers must remain focused is the creation of value for all users of the platform, which strengthens the community, improves its long-term health and vibrancy, and encourages the continual growth of positive network effects.

In this chapter, we’ll consider some of the key issues related to developing and using appropriate metrics for platform businesses, following the life cycle of a platform from startup to maturity. In the startup phase, it is critical to have simple measures to guide decision-making around key questions of platform design and launch, including the design of the core interaction; the development of effective tools to pull users, facilitate interactions, and match producers with consumers; the creation of effective systems of curation; and decisions about how open the platform should be to various kinds of participants.

In particular, firms in the startup phase must track the growth of their most important asset: active producers and consumers who are participating in a large volume of successful interactions. These users and the interactions they engage in are the key to generating the positive network effects that will ultimately make the platform successful. Notice that some of the traditional metrics generally regarded as crucial in the early years of a pipeline business—revenues, cash flow, profit margins, and the like—are largely irrelevant when evaluating a platform during the startup phase.

Once the platform has reached critical mass and users are gaining significant value from the platform, the focus of metrics can shift to customer retention and the conversion of active users to paying customers. This is the phase in which monetization becomes a crucial issue. As we explained in chapter 6, decisions about how to monetize the platform are fraught. Platform managers will need to devise metrics that focus on some of the key issues related to monetization, for example: Which user groups are enjoying the greatest value from platform activities? Which user groups may need to be subsidized to ensure their continued participation? What fraction of the value creation unleashed by the platform is occurring on the platform rather than outside it? How much additional value can be created through services such as enhanced curation? Which groups outside the platform might find value in access to specific user groups on the platform? And most important, how can the platform capture and retain a fair share of the value being created on the platform without impeding the continued growth of network effects? During the growth phase, thoughtfully designed metrics can help platform managers develop accurate answers to questions like these.

Finally, as the platform matures and a self-sustaining business model has been developed, the challenge of user retention and growth requires the platform to innovate. This is the best way to maintain and enhance the business’s value proposition relative to competing platforms. Metrics then must sensitively gauge the ongoing engagement of users and the degree to which they continue to discover new ways to create value on the platform. It’s crucial to measure and track the degree to which both producers and consumers are repeatedly participating in the platform and increasing their participation over time.

Other competitive concerns include attempts by adjacent platforms to drain users and reduce the platform’s comparative advantage, as well as the possibility that participants in the platform (such as extension developers) may create platforms of their own that could eventually pull users away. These, too, call for the development of metrics that will enable platform leaders to recognize such threats and respond to them in time.

At startup companies—whether they operate as pipelines or as platforms—resources are usually stretched. With money, time, and talent at a premium, people find themselves doing multiple jobs, often in domains far from their expertise. In this kind of environment, deciding the categories of information to which to dedicate resources in collection and processing can be both critically important and challenging.

Furthermore, the kinds of metrics that work in a startup context may be quite different from those that apply to a conventional, mature business. Entrepreneur Derek Sivers describes the problem:

Most tools from general management are not designed to flourish in the harsh soil of extreme uncertainty in which startups thrive. The future is unpredictable, customers face a growing array of alternatives, and the pace of change is ever increasing. Yet most startups—in garages and enterprises alike—still are managed by using standard forecasts, product milestones, and detailed business plans.⁵

So what kinds of metrics are most valuable during the startup phase of a platform business? Platform managers should focus on the core interaction and the benefits it creates for both producers and consumers on the platform. To define success or failure for a platform, and to identify how to improve it, there are three main metrics: liquidity, matching quality, and trust.

Liquidity in a platform marketplace is a state in which there are a minimum number of producers and consumers and the percentage of successful interactions is high. When liquidity is achieved, interaction failure is minimized, and the intent of users to interact is consistently satisfied within a reasonable period of time. Achieving liquidity is the first and most important milestone in the life cycle of a platform. Therefore, the most valuable metric in the early months of a platform is one that can help you determine when liquidity is reached. Depending on the precise workings of the platform and the nature of its user base, the formula for this metric may vary.

One reasonable way to measure liquidity is by tracking the percentage of listings that lead to interactions within a given time period. Of course, both the definition of “interactions” and the appropriate time period will vary depending on the market category. On an information and entertainment platform, an interaction might be the click-through that takes a consumer from a headline to a complete story; on a marketplace platform, it might be the purchase of a product; on a professional networking platform, it might be the offer of a recommendation, the swapping of contact information, or a posted response to a question on a discussion page. Any of these interactions would signify a greater degree of engagement by the user and represent the moment when the user has recognized, used, and enjoyed a value unit available on the platform.

On the negative side, it’s important to look for and track the occurrence of illiquid situations. These are circumstances in which a desired transaction is impossible—for example, when an Uber user opens the app and discovers that no car is available. Illiquid situations discourage users from participating in the platform and so must be kept to a minimum.

Note that user commitment and active usage of the platform are the vital metrics of platform adoption, not sign-ups. That’s why our definition of liquidity includes both user totals and the level of interactions occurring. New reports and investor pitches that emphasize impressive raw numbers of platform members can be very misleading and may be a sign that the platform, far from flourishing, is struggling to convert curiosity-seekers into active participants and value creators.

Also note that the most meaningful metrics are comparative ones, which draw helpful distinctions between groups of users or over periods of time (a useful recommendation from Alistair Croll and Benjamin Yoskovitz, authors of Lean Analytics). A good example of an inherently comparative measure is a ratio or rate, which is calculated by dividing one number by another—for example, the ratio of active users, which is calculated by dividing the number of active users by the number of total users, or the rate of growth in active users, which is calculated by dividing the number of new active users by the number of total active users.⁶

A second crucial category of metric for the startup platform is matching quality. This refers to the accuracy of the search algorithm and the intuitiveness of the navigation tools offered to users as they seek other users with whom they can engage in value-creating interactions. Matching quality is critical to delivering value and stimulating the long-term growth and success of the platform. It is achieved through excellence in product or service curation.

As the definition implies, matching quality is closely related to the effectiveness with which products or service offerings on the platform are curated. Users generally participate in a platform with highly interactional intent; they want to find what they’re looking for as quickly as possible. Precision in matching leads to lower search costs for users—that is, they need to invest less time, energy, effort, and other resources in finding the matches they want. Thus, if the platform does a great job of linking users to one another quickly and accurately, those users are likely to become active participants and long-term members of the platform; if matching quality is poor, slow, and disappointing, users will soon dwindle in number, interactions will slow to a trickle, and the platform may be doomed to an early demise.

Of course, it’s necessary to translate the abstract term “matching quality” into a concrete quantity with a clear operational definition in order to make it the basis of a meaningful metric. One way to measure the efficiency of the platform in successfully matching producers to consumers is by tracking the sales conversion rate, which can be expressed as the percentage of searches that lead to interactions.

Obviously, the higher the percentage, the better—but where does the threshold between “poor” and “good” matching quality lie? There’s no single answer that applies to every kind of platform. However, the manager of a particular platform may be able to develop a useful rule of thumb by correlating the interaction percentage for particular users with the long-term rate of activity of those users—say, over a period of one to three months. Calculations like these may enable you to determine, for example, that an interaction percentage of 40 percent appears to represent a significant cutoff point for users of your platform: the majority of users who experience interaction percentage higher than 40 percent during their first week on the platform remain active members for at least three months, while a majority of those with an interaction percentage lower than 40 percent stop participating in activities on the site.

Once you’ve calculated a number of this kind—whether it is 40 percent, higher, or lower—you can use it as a working target that serves as one measure of the health of your site. The daily interaction percentage can be measured, its trend over time can be observed, and improvements to the platform’s matching system can be developed, tested, and evaluated based on changes in this metric.

The third crucial category of startup metric is trust. Trust refers to the degree to which users of a platform feel comfortable with the level of risk associated with engaging in interactions on the platform. It is achieved through excellent curation of participants in the platform.

Building trust, of course, is central to marketplaces, especially those in which interactions carry some level of risk—and in the world of online platforms, where initial connections among users as well as many interactions are conducted entirely in cyberspace, the perception of risk may be even more significant. A well-run platform is one in which participants on both sides have been successfully curated so that users are comfortable with the level of risk involved in engaging in interactions on the platform. As we’ve noted, Airbnb is an example of a player in a high-risk category that has succeeded so far because of its ability to curate its participants successfully. It allows hosts and guests to review each other and has one of the highest review rates among platforms. It also takes additional measures to build trust, including having photographers certify the accuracy of the information contained in a host’s listing. By contrast, Airbnb’s competitor Craigslist has earned relatively low scores on the trust metric and has experienced a number of embarrassing scandals involving apparently sleazy platform users engaged in disreputable, even illegal activities.

These three crucial categories of metrics—liquidity, matching quality, and trust—combine to provide the managers of startup platforms with an accurate picture of the platform’s rate of interaction success and the key factors that contribute to it. As we’ve noted, this measurement is at the heart of the platform’s purpose and plays a central role in determining its ability to create positive network effects.

The specific formulas you use to define the metrics for a particular platform business need to be carefully devised to be appropriate for the kind of platform business involved—the nature of the platform, the types of users, the forms of value being created and exchanged, the variety of interactions performed, and so on.

There are a number of specialized metrics that are potentially valuable for particular platform businesses. You might choose to measure engagement per interaction, time between interactions, and percentage of active users, all of which focus on the degree of user commitment to the ecosystem.

Alternatively, you might choose to measure number of interactions, as, for example, the graphics and design platform Fiverr does. Since Fiverr has a fixed value per interaction—every “gig” traded on the site is priced at five dollars—the sheer number of interactions is a perfectly adequate and complete measurement of the current activity flow on the site.

Other platforms need to develop more sophisticated interaction metrics. Airbnb, for example, tracks the number of nights booked, which is a better indicator of value creation for this platform than simply recording the number of interactions. The freelance work marketplace Upwork measures interaction volume by counting the hours of work delivered by a particular freelancer, which is a key measure of value creation in that ecosystem. In a similar fashion, Clarity can track the duration of a consulting call between an expert and the information seeker.

Platforms whose revenue is based on claiming a share of the value of any interaction—a commission fee based on a percentage of the interaction, for example—may choose to measure interaction capture, which will reflect the value of interactions that occur on the platform. Amazon Marketplace, for instance, uses this metric, tracking the gross value of interactions processed by the platform as a key indicator of its activity level.

Platforms that focus on content creation require different metrics. For example, some measure co-creation (the percentage of listings that are consumed by users) or consumer relevance (the percentage of listings that receive some minimum level of positive response from potential consumers). These metrics focus on interaction quality and reflect the skill with which production is being curated.

Finally, other platforms focus on market access—the effectiveness with which users have been able to join the platform and find or connect with one another, regardless of whether a complete interaction has occurred. Some measure producer participation—that is, the rate at which producers join the platform and the growth of this rate over time. Dating and matrimonial sites often talk about number of women registered, since this metric serves as a useful proxy for the value that other users of the site can expect to receive. In a somewhat different fashion, OpenTable tracks restaurant reservations. These are not the actual interactions, in which restaurants are paid for meals served (information not readily available to the platform), but they serve as a fairly accurate proxy for the value created.

The three key factors of liquidity, matching quality, and trust remain crucial to measuring the health of virtually any kind of newly launched platform. But as you can see, specific characteristics of a particular platform may dictate the need for additional, more specialized measurement tools. The variety and range of metrics that may be suitable during a platform’s startup phase is limited only by your ingenuity and the nature of the activities occurring in your burgeoning ecosystem.

The metrics that best measure the number and quality of interactions in your ecosystem will change over the life cycle of the platform, and it’s critical to identify points at which these transitions occur. Companies often make the mistake of clinging to metrics that their business has outgrown. Identifying and vetting the core metrics that are most relevant to the decisions you face today is important at every point in the platform’s development.

For example, once the platform reaches a critical mass of users, new issues arise. Managers must still ensure that the core interaction is creating value and that the inflow of engaged users exceeds the outflow so that the platform is still growing. However, as growth continues, the platform must monitor the change in size of the user base over time. In particular, platform managers will want to work to ensure balance on the two sides of its market. This balance can be monitored by calculating the producer-to-consumer ratio, with an adjustment to include only active platform users—those who’ve engaged in interactions on the platform at a specific minimum rate of frequency that you consider appropriate. Experience shows that this ratio is a crucial factor in the rate of interaction success achieved by the platform.

Consider the core interaction that the dating website OkCupid is facilitating: introductions between men and women. As we noted in chapter 2, one of the critical things for this platform to manage is the access of straight men (who may be considered “consumers” in this context) to straight women (who play the role of “producers”).* As a result, OkCupid tracks the ratio of straight women to straight men, and platform managers work hard to adjust that ratio when it diverges from the level they deem optimal. They manage these adjustments by asking users to rate the attractiveness of those on the opposite side of the platform.⁷ The website then introduces a filter to reduce the number of men who can participate in the platform by seeing women’s profiles—especially women who are rated as particularly attractive.⁸ In this way, the OkCupid platform is helping to maintain positive network effects and fostering market liquidity by avoiding an imbalance that might otherwise alienate a segment of its female users. Continually measuring and monitoring the male–female ratio makes this maintenance possible. In similar fashion, the freelance platform Upwork focuses on keeping the number of freelancers proportional to the number of job postings, since a surfeit on either side causes participants to leave.

For a traditional two-sided platform with producers on one side and consumers on the other, it is best to find ways to calculate the value of each user type. In Lean Analytics, the entrepreneur/author team of Alistair Croll and Benjamin Yoskovitz provide a useful illustration of metrics for a two-sided platform, which we adapt below.⁹

On the producer side, the platform should monitor figures that include the frequency of producer participation, listings created, and outcomes achieved. The platform should also monitor interaction failure—the percentage of cases in which interactions, such as sales, are initiated but fall through for some reason. This is a crucial metric that many platform managers overlook. If users are being retained but the rate of successful interaction is falling, there is a serious problem.

It’s especially important to monitor instances of producer fraud—for example, the failure of a producer to describe a product offering accurately or to deliver it in a timely fashion. Producer fraud is, of course, a particularly egregious, painful, and costly form of interaction failure. Examination of the characteristics of users and interactions repeatedly linked to fraud may be used to create predictive models that can help the platform prevent future fraud.

Combining all these forms of data, the value of a producer can be calculated using traditional lifetime value (LTV) models used in many kinds of businesses. These models capture the mechanism by which repeat producers provide recurring platform revenues without incurring additional acquisition costs—that is, expenses incurred by the platform in attracting and engaging these producers. Because repeat producers are especially profitable to a platform, well-managed platform firms will work hard to create active repeat producers, just as subscriber-based services like magazines and cell service providers work to keep the rates of subscriber turnover (or churn) as low as possible.

On the consumer side, the growing platform should monitor the frequency of consumption, searches, and rate of conversion to sale (the percentage of click-throughs that result in completed interactions). This information, along with the likelihood of repeat interactions, provides the data necessary to calculate each consumer’s LTV. Once both producer and consumer LTV measures have been created, the platform can run experiments in an effort to impact the critical determinants of LTV—churn rate, for example.¹⁰

Most of today’s successful platform businesses have programs designed to encourage loyalty on the part of the most valuable active users and to discourage those who are less valuable. If you’ve ever had a platform like Facebook or LinkedIn ping you with an invitation to return to the platform after a falloff in your usage, you’ve been targeted by such a program. Similarly, Twitter has introduced the “popular in your network” feature to alert you to content that might be particularly relevant even if you have not subscribed to those authors’ feeds—another activity-building program that is driven by metrics and designed to stimulate more interactions around users with a proven track record of value-creating activity.¹¹

A critical variable from the startup phase that remains highly relevant during the growth phase is the interaction conversion rate—that is, the percentage of searches or queries that result in interactions. Well designed and consistently monitored metrics focusing on the sales conversion rate can help platform managers develop smart strategies that will enhance the platform’s continued growth—as when Airbnb introduced its professional photography service after discovering that high-quality photos increase property rental rates.¹²

Interestingly, Airbnb has also discovered that its best source of hosts is people who have been guests. Consequently, it is now working hard to convert consumers on its platform into producers. In this case, the side switching rate—the rate at which people convert from one type of user to another—offers an important metric that the platform can use to track the health of its user base and to maintain balance across its network.

New metrics are continually being devised by platform managers based on their specific objectives and interests as well as the unique characteristics of their users. Haier Group is a rapidly growing manufacturing company based in Qingdao, China. It is currently building a platform to connect its customers with the design and production teams, both inside and outside the organization, that create the products, which include home appliances and electronics. Haier’s CEO, Ruimin Zhang, spoke with the authors about a unique metric that the company is eager to capture and use—namely, the distance between consumers and producers.¹³ In this case, the word “distance” is metaphorical, not literal; it refers to the frequency of direct interaction and the size, reach, and influence of the social networks that connect producers of Haier products to their users.

To measure this distance, Haier has devised metrics based on interactions on WeChat, a social instant messaging and photo sharing tool developed by the Chinese company Tencent. The goal: to minimize the distance between Haier and its customers, thereby improving the fit between products and consumer needs, enhancing the company’s innovative capacities, and making its marketing and promotional efforts less costly and more effective.

As CEO Zhang pointed out to us, the size of a company’s advertising budget might be viewed as a reflection of the distance between the company and its customers. For example, the annual brand value report issued in 2013 by the consulting firm Interbrand noted that Google’s advertising budget is just a tiny fraction of Coca-Cola’s. The likely reason: Google is deeply integrated into people’s lives through its many productivity and social applications, giving it constant user feedback that Coca-Cola doesn’t receive.

Based on analogies like this, Haier’s leadership team hypothesizes that a reduction in its user distance measure may improve its product design, customer service, and marketing efficiency. Thus, a seemingly abstract metric like user distance may have a highly practical, dollars-and-cents impact on your bottom line.

Once a platform business has moved past the phases of startup and early growth, new challenges and issues emerge. Eric Ries, the writer and entrepreneur known for pioneering the “lean startup” movement, emphasizes that, for the mature company, incremental innovation and metrics must be closely related to each other. “When making improvements to your product,” Ries observes, “the only arbiter of whether or not it was successful is the metrics. And, when you are implementing an improvement to your product, you should be testing that improvement against a baseline.”

Somewhat in line with Ries’s thinking, Amrit Tiwana, a professor at the University of Georgia, suggests that metrics suitable for information technology platforms that have reached the maturity phase should meet three major requirements: they should drive innovation, have a high signal-to-noise ratio, and facilitate resource allocation.¹⁴

First, let’s focus on the role of metrics in driving innovation. In order to remain vibrant, a platform must be able to adapt to the needs of its users and to changes in the competitive and regulatory environment. One way for a platform to identify necessary adaptations is by studying the extensions provided by developers. These innovations may represent functionalities missing from the core platform, which the platform may choose to absorb. For example, during the era of the desktop computer, Microsoft Windows absorbed a number of applications that were once provided by stand-alone companies, such as disk defragmentation, file encryption, media playing, and more.¹⁵

Cisco has followed the same absorption strategy in the router business, where it operates a platform known as the Cisco Application Extension Platform. The Cisco AXP (as it is called) is a Linux-based platform that allows third-party developers to create applications that work on Cisco routers, providing new capabilities that Cisco customers find useful—for example, enhanced security measures and customized monitoring systems. When we asked Cisco’s chief technology officer, Guido Jouret, how the company decided which functions to bundle into the Cisco AXP, his answer was illuminating:

The issue is to embed into the platform multiple independent solutions to the same problem. Then this becomes common for everyone else. It’s a question of timing. If you do it right away, your ecosystem is scared that you’ll cannibalize their cash cows. If one provider builds a particular functionality, you don’t want to co-opt it. But if a whole slew of them have [developed the same capability], then competition reduces benefits anyway, and you can fold it in.¹⁶

To enable this strategy, Cisco employs metrics that seek out instances in which the same capability is provided across multiple industry verticals—health care or the automotive business, for example. That is a sign that the platform is missing important features that should be part of the next round of continuous platform innovation.

A platform may also choose to innovate when features provided by third parties become a large part of the overall value enjoyed by users. As we saw in chapter 7, this helps to explain Apple’s 2012 introduction of Apple Maps in response to the enormous popularity of Google Maps.

Some kinds of platforms need still other customized metrics during their maturity phase. These include labor platforms such as Upwork, data platforms such as Thomson Reuters, connection platforms such as Skype, and platforms that connect machines such as GE’s Industrial Internet. Although these are distinct platform types with disparate needs, they all face the challenges of facilitating a core interaction, measuring the drivers of value, and innovating to maintain the platform’s ability to produce significant value for users.

The metrics dashboard you develop for your platform can be quite complex, allowing you to get a real-time glimpse of activities at a very fine level. However, simplicity is a virtue when developing metrics for your platform business. Overcomplex metrics make management less effective by introducing noise, discouraging frequent analysis, and distracting from the handful of data points that are most significant.

At one time, oDesk (now known as Upwork) had so many metrics (measuring job postings, registered workers, service variety, and many other factors) that one board member complained, “You’re over-measured and under-prioritized.” Having learned from this mistake, Gary Swart, former CEO of oDesk, writes eloquently about the need for highly focused metrics, especially in the critical early period of a startup:

As a business leader you need to figure out the metric that matters most for your company and understand that the more you measure, the less prioritized you’ll be. Don’t fall into the trap of trying to measure everything. What I’ve learned is that in the early days, what matters most is having customers who love and use your product. Figure out the one or two best measures to determine this.¹⁷

Lean startup guru Eric Ries echoes the need to be selective in the design and use of metrics. In particular, he cautions against what he calls “vanity metrics,” such as total sign-ups—a relatively meaningless statistic that often increases even as the volume of interactions is flat or actually declining. Vanity metrics fail to indicate accurately whether the business is really achieving critical mass or the liquidity it needs.

Instead, Ries suggests, “you should make sure your metrics meet the ‘3 A’s test’ where your metrics are actionable, accessible, and auditable.” They must be actionable in that they provide clear guidance for strategic and managerial decisions, and in being clearly related to the success of the business. They must be accessible in that they are comprehensible to the people who gather and use the information. And they must be auditable in the sense that they are real and meaningful—based on clean, accurate data, precisely defined, and reflecting the reality of the business as perceived by users.¹⁸

In the end, the most important metric is a simple one: the number of happy customers on every side of the network who are repeatedly and increasingly engaged in positive, value-creating interactions. The real question, which you should never lose sight of, is: are people happy enough with the ecosystem to continue participating in it actively? No matter how you end up designing the metric dashboard for your specific platform business, it should ultimately serve to accurately measure the answer to this key question.

TAKEAWAYS FROM CHAPTER NINE

Since the value of a platform is derived primarily from network effects, platform metrics should ultimately seek to measure the rate of interaction success and the factors that contribute to it. Interaction success attracts active users and thereby enhances the development of positive network effects.

During the startup phase, platform companies should concentrate on metrics that track the strength of characteristics that enable core interactions on the platform, including liquidity, matching, and trust. These characteristics can be measured in a variety of specific ways, depending on the nature of the platform.

During the growth phase, platform companies should focus on metrics that are likely to impact growth and enhanced value creation, such as the relative size of various portions of the user base, the lifetime value of producers and consumers, and the sales conversion rate.

During the maturity phase, platform companies should focus on metrics that drive innovation by identifying new functionalities that can create value for users, as well as metrics that can identify strategic threats from competitors to which the platform needs to respond.

*We recognize the unpleasant implications of this language. It reflects the currently prevailing dynamic of many male–female dating interactions in U.S. society at this time, including the fact that most online dating sites find it easier to attract male participants than females. Thus, females are “in demand” in a way that is analogous to the demand for highly-sought-after products on an auction site like eBay. As social norms evolve in the direction of greater gender equality, we hope and expect that this dynamic will also evolve, with implications for the effective management of dating platforms.