Thoughts on Growth — May 11, 2018

Thoughts on Growth is Reforge's newsletter of must-know updates and perspectives in growth. By subscribing, you'll join a few thousand PMs, marketers, UX folks, engineers and analysts at today's top tech companies. Check out today's edition below.

1. Freemium or free trial for B2C monetization and retention



A study by Pandora showed that serving more ads on its free trier increased conversion rate to its premium tier. But, that lift was canceled out by the drop in user engagement and retention that those ads caused.

Other takeaways:

  • Serving just one additional ad per hour resulted in a 2% drop in average session listening time.
  • The number of days a user listened also dropped 1.9%.
  • Each additional ad per hour lead to a 0.14% increase in upgrades to premium (users in older age brackets converted at a higher rate).
  • But advertising revenue lost from users churning outpaced revenue gained from upgrades.
  • For every user that upgraded, three churned.


Meanwhile, Spotify has been adding more features to its free, ad-supported tier to increase listener engagement in order to grow paid subscriptions.

“You can put in a free user and 40 percent come out as a payer,” Gustav Söderström, Spotify’s chief R&D officer told Wired.

That is much higher than typical freemium conversion rates, which are often in the low single-digits.

To add another data point, Apple Music, which has a free trial (not freemium) model, has been growing its U.S. subscriber base by about 5% a month, versus Spotify’s 2% per month. This puts Apple Music on pace to pass Spotify in paid U.S. subscribers.

BRIAN BALFOUR CEO @ Reforge, former VP Growth @ Hubspot:

Keep in mind that freemium is an acquisition model, not a monetization model.

Freemium is a way to lower friction and enable virality or some other form of acquisition model.

Apple doesn't need it because their acquisition comes from other sources (selling the iphone, apps marketplace, etc). That's why Free Trial works for them; they already own the user.

Spotify (and Pandora) needed freemium because they didn't already own the user. They needed a low cost way to acquire a bunch of users.

In the Pandora study, one conclusion was that more friction (ads) = more subscribers. This is only true if you're thinking of freemium as a monetization model, which it isn't.

Spotify is going down the track that more engagement in free tier leads to more acquisition of new users and longer retention of those free users. That combination of greater acquisition, higher engagement and longer retention eventually leads to more upgrades.


2. The compounding effects of “negative churn”

  Image data from David Skok

Image data from David Skok

DAVID SKOK, General Partner @ Matrix Partners and Board Member @ HubSpot on three key ways to get to negative churn:

  • Expand revenue from your current product. This is best done by having a pricing model that increases the pricing according to some usage metric that will grow over time.
  • Up-sell customers to a more highly featured version of your product.
  • Cross-sell customers to purchase additional products or services.

While not a hard and fast rule, my usual recommendation is to split the sales organization into hunters that chase deals with new customers, and farmers that work on expanding the revenue from existing customers.

To reduce churn:

  • Call your customers to find out why they're canceling, or not renewing.
  • Measure and score customer engagement, to get an indicator of which customers are getting good use out of the product, and therefore not likely to churn, and which are at risk, and need intervention.
  • Figure out what features make your product sticky, then use measurements of customer engagement to see which customers are not using those features. Those are the customers most at risk of churning.
  • Allocate your best reps to the job of saving customers, because it requires your best sales skills to achieve good results.
  • Test a longer term contract up front (usually 6 or 12 months).
  • Look for other factors that correlate with churn, like customer segment, lead source, etc, in order to inform revisions on marketing or product.


3. The perils and perks of building on a platform


ANDY COOK, co-founder & CEO @ Tettra and formerly PM @ HubSpot:

In 2006, the average number of choices a consumer had when choosing a product was four or five. In 2016, that number has spiked to approximately 15 vendors. The key to winning is being the best at distribution and showing up where your users already hang out.

Connecting to existing platforms like Slack makes it easier for your users to access your product and helps you grow. People don’t want another login to remember or interface to learn. If you’re early enough to a platform and build on top of it right, hopefully eventually you’ll grow big enough to become a platform yourself.


  1. Platforms could go bust or cut off your app access.
  2. Platforms will move underfoot. Policy or algorithms changes could impact how your product or content gets distributed.
  3. Platforms have your product usage data and can decide to clone a successful product.
  4. Should you ever wish to leave the platform, you'll likely be starting (almost) from zero, and if you've been relying on platform "built-ins" to juice discovery, distribution and retention, it will be a tough transition to handling those on your own.


  1. Distribution. There's potential to tap into existing user base of the platform.
  2. Product- or feature-driven retention and engagement (the platform already has features designed to reengage users, so you get to ride those coattails — like when you publish on Medium instead of on your own blog, and get to take advantage of all the follower notifications and email marketing that Medium already built. This also applies to discovery / acquisition, as well as to retention and engagement)
  3. Reduces friction to building a habit. (Your potential / target users already have a habitual relationship with the platform, so instead of carving out a new behavioral path, you get to just tap into what they're already used to via their platform usage).


4. Paid acquisition's local maximum & what to do about it



Product marketing campaigns decrease in efficacy and increase in cost due to:

  • Novelty effects - the message, once fresh and offering new customer value, becomes stale
  • Audience saturation - after exhausting the core demographic, campaigns must expand to a broader audience
  • Competitive dynamics - competition copycats your product and messaging and you're no longer unique

ANDREW CHEN General Partner @ Andreessen Horowitz on our addiction to paid marketing:

The new generation of ad platforms makes it possible to scale revenue to new heights, but without profitability. Make sure you don't get addicted.

Addiction to paid marketing can get you into a local maximum. It's much harder to fix the underlying issues - creating real moats, product differentiation, doing deeper adtech integrations. Easier to just spend more and push the LTV window from 9 months to 12 to 18.

There's a few scenarios where paid marketing is justified, but it's situational. If your product has network effects that kick in after an activation point and really scale, you can use paid to help bootstrap that. Facebook uses paid to build out new regions, for example.

If you are really going to invest a ton of time from engineering/growth to integrate with all the APIs, try out a ton of things algorithmically, then you can develop a lasting edge. I've heard Wish does this well, but it's not common.

Build out-there channels. Fix churn and frequency. Don't congratulate yourself too early. And, calculate LTV/CAC correctly.



Thoughts on Growth is Reforge's newsletter of must-know updates and perspectives in growth.