Many Falkoners, including our co-founder, are early Dropbox people. Given product-led growth (PLG) is such a hot topic right now, we want to share things we learned from our experience at Dropbox—as with all learning, this includes things that worked AND things that didn’t work / were gaps.
The lessons we learned range from how to create growth loops to monetization techniques to how cultural friction can impact growth.
We had a hypothesis that users who were approaching their quota limit were most likely to convert, making them our best product qualified leads. So, we sent them many marketing messages with incentives to convert. We got an initial lift in conversion, but it plateaued. It took several iterations to figure out that the next conversion lift would come from combining two metrics: 1) customers who were approaching their quota limit and 2) who had also been using their quota at a high rate, signaling they were going to hit their limit soon. Simply put, if a user is at 90% of their quota but it’s taking them six months to use up 1% of quota, they’re actually pretty far off from hitting their limit.
The lesson: Form many hypotheses about what metrics signal conversion or expansion readiness. Then validate them with statistical analysis to truly identify the leading indicator metrics. Use the metrics to spot conversion/expansion ready users. Repeat this process regularly to unlock more opportunities for conversion lifts. The idea of looking for more opportunities brings us to the next big lesson.
Our growth team had a rich set of user segmentation criteria based on product usage around specific features. They used the output for highly targeted and personalized outreach campaigns. For instance, users on iOS converted better than ones on Android. However, users on both iOS and MacOS converted even higher. The value of the service for iOS- or Android-only users was primarily about photos, while the value for iOS + MacOS was also about synchronizing data across devices. Targeting these fine-grained segments with incentives that unlocked the specific value they got from the service was key to conversion.
The lesson: Doing coarse segmentation like “this set of customers is ready for conversion” is necessary, but insufficient. You need specific usage signals so you can create marketing campaigns and incentives that appeal to different users within the broad set of conversion-ready users.
From the lesson above, highly personalized outreach increases conversion. However, managing hundreds of different user lists based on segmentation criteria is impossible as a manual task. That’s why an automated workflow is essential. Downloading spreadsheets just doesn’t cut it. In our case, the growth team had a homegrown solution that provided rich customer segmentation and automated marketing outreach. Unfortunately, the marketing and sales teams used two systems (Marketo and Salesforce) and the systems were never connected in a seamless way. They did periodic data dumps and therefore were unable to execute on personalization at scale.
The lesson: Automation isn’t only about saving time. Personalization at scale is impossible without it.
The lesson above alluded to this and it’s an important one to state explicitly: when teams aren’t working together, known opportunities to drive monetization can go unactioned. Dropbox was an engineering-led organization with a strong growth arm within the product team. Even after traditional marketing and sales teams formed, their tools and culture remained distinct from the product organization. This meant that many of the insights we generated within the product organization didn’t get acted upon, allowing incredible monetization opportunities to fall through the cracks. This showed up in tactical and strategic ways. Here are three illustrative examples.
Example 1: Hidden segment that was never monetized
In our fraud work, we found a great way to identify corporate users who were hidden and not being targeted because they were using personal email addresses. We matched billing information and billing address. If we saw 30 users with the same last 4 numbers of a credit card or the same billing address, it signaled either fraud or corporate users who weren’t being tracked or targeted as corporate users. However, there was no way for us to push this information to marketing and sales (different teams, no shared meetings or overlaps) or to push the information to their tools/systems of action like Salesforce and Marketo. Also, since the communication across teams wasn’t strong, we were unable to identify the size of the opportunity relative to other opportunities marketing and sales were pursuing.
Example 2: Team-specific understandings of the ideal customer profile
Strategic planning was based on market research, qualitative studies, and periodically through data. Different teams (marketing, sales, product, and growth) formed distinct understandings of the ideal customer profile coming from these inputs. However, for companies that have a lot of non-paying customers (like Dropbox did), the real answer to the ideal customer profile resides in product usage data. Every team should have shared an ideal customer profile based on product usage among the best segments. If teams were working together and had access to the same data, the types of prospects to acquire via traditional GTM channels would have been much more clear.
Example 3: Hesitancy to update pricing based on value metrics
As product teams rolled out new features, we could see clearly which features were driving the most engagement and value for users, but our input into pricing updates was minimal. For instance, “offline folder on mobile,” which allowed users to listen to music on their phone even when they weren’t internet-connected, was a highly valuable feature for power users. The product team knew about this signal for a long time, but the pricing team was slow to react to the information that power users would pay for the feature. It took a lot of time, effort, and unnecessary friction before offline folders were put behind the paywall. Similarly, syncing files across multiple devices was a high-value capability. The product team knew that highly engaged users had many devices and attaching pricing to syncing them would be lucrative. This, too, is now factored into Dropbox’s pricing, but was a challenge for the same reasons.
The lesson: Product usage data can power many strategic and tactical decisions, from prospecting to pricing. For this data to speak, however, companies need explicit focus on overlapping meetings, synched goals, and shared data access across teams.
This is an oldie but goodie and we share it because creating a win-win referral program within your product is a highly cost effective way to do expansion. We see many companies offer users cash or discounts in exchange for referrals. While users would appreciate cash savings or cash in pocket, this benefit is a win for them but not really a win for our service since it doesn’t encourage anything beyond a referral link. With Dropbox, the “prize” for referring the service to a coworker or friend was unlocking additional quota. The user got something valuable and so did we–not just a referral link, but increased usage from the referrer.
The lesson: Come up with a referral “prize” or incentive for users that they perceive as beneficial, and that also drives good engagement behaviors for your company.
Want a peek at the Falkon PLG engine?