Can data tell you which channel will drive the next round of growth - product, sales, marketing?

One of the toughest questions facing growth managers and growth companies is this - which channel will drive the business’s next round of growth? Product? Sales? Marketing?

Such question is much easier to answer in the early stage of a company. Signal from zero to one is easy to detect. There is a growth playbook to follow, with established metrics for each big bet - mobile, SEO, social, sales.

Following the initial high growth, most companies enter an extended period of stable growth, during which proven growth initiatives are scaled and continue to deliver incremental benefits. As the company grows in size and organizational complexity, big initiatives are distributed to functional priorities. This is when companies introduce company KPIs as well as functional KPIs, in order to make sure individual initiatives and their joint impact are both on the right track.

Then, came the first speed bump. The slowdown is usually observed through a key business indicator. And it often develops like a chronic pain rather than an acute illness. When the signal is first noticed, it is hard to tell whether it is a temporary glitch or a persisting trend. When the numbers continue to soften; and other related metrics also start to show fatigue, the symptom is confirmed. The logical next steps are to diagnose; and to find cure.

I often joked this is the time companies start seriously building out an analytical function. Now the business is more complicated. There is more data. Both the internal and external expectations are higher. The margin of error is smaller.

The ideal scenario will be for the newly minted analytical team to have all the necessary data, the right analytical tools, and the clear understanding of the business drivers to come up with definitive diagnosis. That rarely happens. More often they find themselves struggling with distributed intelligence that does not paint the full picture of the intricacy between business drivers; or a makeshift analytical infrastructure that has been making room for other more important revenue generating engineering development; or the religious attachment to the entrepreneurial instinct that takes data as the proof of a decision, not the decision itself. In a perfect storm, they face technical shortages and cultural stubbornness; while the pressure of showing continuous strong growth breathes down everyone’s neck.

While it can be one of the most difficult times for the business, it can also be a defining moment. Just think about all the established companies who still launch successful products and drive sustainable high growth. How do they do it?

One thing they did right is to see the jungle, the complex ecosystem their businesses operate in. It is not possible to remove the complexity or pretend it does not exist, but it is possible to make better map. Such map is data. In today’s business, each function generates huge amount of data - all the product interactions, marketing touch points, social engagement, etc. The problem is usually not the lack of data, but rather the linkage between them. For example, without connecting the Salesforce data to Marketing, we would not know an account acquired by the Sales team actually started her journey interacting with marketing campaigns. The big data technology and lower storage cost allow us to go deep in tracking everything a user does. But it still requires forward thinking data architecture and thoughtful implementation to build a data foundation which provides both big picture insight as well as spotlight intelligence.

There are many ways to draw a map that represents the same underlying data. Take the evolution of subway maps for example. Eighty five years ago, Harry Beck’s London underground tube map defined the transport network map design ever since. By forgoing the precise geo locations, the map provides passengers intuitive understanding of the system, thus allow them to digest the otherwise complex information within reasonably short amount of time. The equivalent of Harry Beck’s map in business is the set of metrics, or KPIs, that help companies to keep close tabs on the key business drivers.

One of the best examples of this is Facebook’s engagement metrics. Even in its early days when acquisition tends to be the only focus, Facebook understands the stickiness of its site, and later its apps, is the key to users’ time spent and the amount of information collected. Both increases the opportunities for monetization. Though Facebook is sophisticated and advanced in many other data capabilities, this relatively simple engagement metric has remained a key measurement.

Its longevity also speaks to its sensitivity to business levers - initiatives Facebook launched through the years. If a company introduces a new business driver and the KPI does not move, it can mean two things: a) the initiative is not successful; or b) the measurement is wrong. You can see how misunderstanding of these two can impact both short-term investment and long-term strategic decisions.

In practice, such data and analytical design principles are reflected in a few best practices.

• Data is a shared responsibility between Engineering and Analytical teams. By including the analytical team early in the data design process, Engineering team can build the most relevant data that can be utilized to produce insights right away. Analytical team share the burden of explaining the complex questions they try to answer. In most cases, such cross-functional partnership can be greatly improved by a Data PM, who bridges the knowledge gap between technical and business domains; who can also play an arbitrator role ensuring ongoing data quality.
• Different business functions will continue generating different data points. The connectivity among them need to be identified early on so all the data collection will have such linkage embedded. For example, the user ID is the key to link product, marketing and sales data, as well as data across devices. Data elements shared by functions, such as geo location, should be collected and processed using a consistent logic. Small details in the upstream data pipeline matter great deal in the downstream data consumption. If inconsistency exists in the data sources, the burden of imposing consistency will fall on the analytical team, thus increases the time required to produce insights.
• Metrics are mapped to key business drivers. For each key business levers - product, marketing, sales, pricing, customer services - there should be a corresponding set of metrics. These metrics evaluate the business driver in its “isolation”. As all business drivers essentially affect the same user, their compound impact is measured by a set of company KPIs. During the market expansion, acquisition through product appeal can be measured by new visitors and visits coming from non-marketing traffic sources. Marketing user acquisition campaigns can be measured by attributed conversions such as sign-up. Sales measures the market penetration and deal sizes. Overall, the company measures the total number of new users, the year over year growth, and their revenue contribution. As the business shifts its focus to monetization, additional metrics are introduced. Product now monitors the completion of the purchase funnel. Marketing optimizes their campaigns based on users’ lifetime value. Sales differentiates account management by revenue tiers. Overall, the business measures the retention rate, the quality of user segments, and the distribution of revenue between new and existing users. Using A/B testing and other control/test methodologies, it can also monitor how the metrics respond to the tuning of business levers.
• As Kevin Kelly observed is his book Inevitable, today’s “new” technologies are the result of long-term, accelerating forces started in the past. What drives today’s growth - the seeds were planted years before. What is happening now will influence the business in the years to come. It is unlikely we will be able to change the business trajectory overnight. Silver bullets are hard to come by. The measurement of business fundamentals should carry such continuity. As the metrics evolve to reflect new lines of businesses, existing metrics continue to be used to link the past and future into a coherent growth story.