If some forecasting was built into the dashboard, I think you'd see lower churn of newsletter creators in the first and second year. They need to be able to visualize what their creativity could become. I may be wrong but my impression is there is a blind spot.
The majority of newsletter writers don't seem to take a long-term view of what paid subscriptions. Plus sponsored ads with flat fees can become in terms of a side gig.
If Substack did what Paved does, and took 15% instead of 30%. What a different business model this would all be for you and us.
Yes, totally agree on the retention point. It's pretty different, but it reminds me of how Wealthfront shows what your investment today will turn into decades into the future. It's strong encouragement to save / invest today, even if the immediate returns on nominal.
Yea that approach makes a ton of sense essentially applying a % of the total. Yea makes sense, especially as you said, with the retention behaviours being similar. Thanks a ton! Excited to read the rest of the articles on your page
This is a really great article. I'm wondering how it differs (or not) for instances when forecasts need to be at a product level. For example a company offers several categories of products and wants to forecast the subscription revenue that will be generated from each category. Would it be just doing this but for each product and them summing it all together? or is there a way to think about it that'll make it less cumbersome
You can make these models as detailed as you want, but that creates more overhead and work. Also, the lower the volume of data, the harder it is to forecast.
One example - in the streaming world, we'd often blend together monthly subscriptions across our primary tier (90% of paid subs) and our ultra-premium tier (10% of paid subs). The retention was close enough to where it didn't really make sense to forecast these two tiers separately. Then we just applied a weighted-average price to the blended paid subscribers — (90% x $8) + (10% x $12) = $8.40 — to get our revenue forecast.
It's really up to the forecaster how detailed you want to get things. I usually encourage keeping it as simple as possible.
Great concept and model, Reid. I’d love to take a look at the actual file if you wouldn’t mind sharing a copy. Also, curious why you assumed in the “% of subs remaining after“ chart that 100% of annual subscribers would be retained after year 1. Why wouldn’t some churn after year 1?
For the annual subscribers, that chart/table is a bit funky because we're comparing the percent of monthly vs. annual subs that remain after the first year. In that view, *all* annual subs will technically remain after the 1st year as they are basically pre-paying for the first year.
We want to know what share of them will renew into year 2, which we factor into our forecast for month 13. This is also when we would expect payment for the 2nd year for annual subscribers, factoring into our cash forecast.
Hope that makes sense, or maybe it will help to poke around the model a bit. Let me know if you have any questions!
Hey Rich, thank you for the questions & sorry for being slow to reply.
Below is a link to a Google Sheet with the forecast model. I haven't gone through this with a fine-tooth comb, but it should at least give you a sense of the structure of these models.
Great article Reid...I'm largely just here to ogle that eye chart...what a beaut! And thanks Tony for giving such an intimate peek into your business...
...curious if you have any thoughts on how to model the growth opportunities for the business expansion components (hiring a growth lead vs. starting a pod vs. buying some programatic ads, etc.)?
I always find it best to try to put it together like a puzzle. So the first step would be to figure out the cost. If you're hiring a growth lead, how much will that cost, and what kind of budget are they going to ask for (above & beyond what you're paying them to do the marketing)?
Once you have a sense of cost, then you can see what you have to believe for to recoup that cost. If it's a $10,000 investment, and your CLV is $100, you would need 100 *incremental* new paid subs for that to be a good investment. Related to the post above, you also need to make sure you have an extra $10,000 in cash flow to pay for the investment.
If you're comparing the different opportunities, that's probably a bit more of a gut call, and something that would be unique to each scenario. For some folks, starting a podcast may be the best investment / priority; for others, hiring a growth lead may be a better decision. Tricky stuff... but often, following your intuition (to an increasing degree over time) is usually a pretty good idea.
I wish these were somehow part of what we share with / show writers; these are all so good and seem really useful and you could see versions of them as being “official Substack documentation” or whatever!
If some forecasting was built into the dashboard, I think you'd see lower churn of newsletter creators in the first and second year. They need to be able to visualize what their creativity could become. I may be wrong but my impression is there is a blind spot.
The majority of newsletter writers don't seem to take a long-term view of what paid subscriptions. Plus sponsored ads with flat fees can become in terms of a side gig.
If Substack did what Paved does, and took 15% instead of 30%. What a different business model this would all be for you and us.
Yes, totally agree on the retention point. It's pretty different, but it reminds me of how Wealthfront shows what your investment today will turn into decades into the future. It's strong encouragement to save / invest today, even if the immediate returns on nominal.
Thanks Michael! 🙏
Yea that approach makes a ton of sense essentially applying a % of the total. Yea makes sense, especially as you said, with the retention behaviours being similar. Thanks a ton! Excited to read the rest of the articles on your page
Thank you! 🙏
This is a really great article. I'm wondering how it differs (or not) for instances when forecasts need to be at a product level. For example a company offers several categories of products and wants to forecast the subscription revenue that will be generated from each category. Would it be just doing this but for each product and them summing it all together? or is there a way to think about it that'll make it less cumbersome
Thank you, Efe!
You pretty much nailed it.
You can make these models as detailed as you want, but that creates more overhead and work. Also, the lower the volume of data, the harder it is to forecast.
One example - in the streaming world, we'd often blend together monthly subscriptions across our primary tier (90% of paid subs) and our ultra-premium tier (10% of paid subs). The retention was close enough to where it didn't really make sense to forecast these two tiers separately. Then we just applied a weighted-average price to the blended paid subscribers — (90% x $8) + (10% x $12) = $8.40 — to get our revenue forecast.
It's really up to the forecaster how detailed you want to get things. I usually encourage keeping it as simple as possible.
Thanks again for dropping the comment!
Great concept and model, Reid. I’d love to take a look at the actual file if you wouldn’t mind sharing a copy. Also, curious why you assumed in the “% of subs remaining after“ chart that 100% of annual subscribers would be retained after year 1. Why wouldn’t some churn after year 1?
For the annual subscribers, that chart/table is a bit funky because we're comparing the percent of monthly vs. annual subs that remain after the first year. In that view, *all* annual subs will technically remain after the 1st year as they are basically pre-paying for the first year.
We want to know what share of them will renew into year 2, which we factor into our forecast for month 13. This is also when we would expect payment for the 2nd year for annual subscribers, factoring into our cash forecast.
Hope that makes sense, or maybe it will help to poke around the model a bit. Let me know if you have any questions!
Hey Rich, thank you for the questions & sorry for being slow to reply.
Below is a link to a Google Sheet with the forecast model. I haven't gone through this with a fine-tooth comb, but it should at least give you a sense of the structure of these models.
https://docs.google.com/spreadsheets/d/1gttjgjeldR0TTP2o-BzRGoUH35ZP0AUaYiTFtebPnpA/edit?usp=sharing
Great article Reid...I'm largely just here to ogle that eye chart...what a beaut! And thanks Tony for giving such an intimate peek into your business...
...curious if you have any thoughts on how to model the growth opportunities for the business expansion components (hiring a growth lead vs. starting a pod vs. buying some programatic ads, etc.)?
I always find it best to try to put it together like a puzzle. So the first step would be to figure out the cost. If you're hiring a growth lead, how much will that cost, and what kind of budget are they going to ask for (above & beyond what you're paying them to do the marketing)?
Once you have a sense of cost, then you can see what you have to believe for to recoup that cost. If it's a $10,000 investment, and your CLV is $100, you would need 100 *incremental* new paid subs for that to be a good investment. Related to the post above, you also need to make sure you have an extra $10,000 in cash flow to pay for the investment.
If you're comparing the different opportunities, that's probably a bit more of a gut call, and something that would be unique to each scenario. For some folks, starting a podcast may be the best investment / priority; for others, hiring a growth lead may be a better decision. Tricky stuff... but often, following your intuition (to an increasing degree over time) is usually a pretty good idea.
Thanks for the response amigo!
I wish these were somehow part of what we share with / show writers; these are all so good and seem really useful and you could see versions of them as being “official Substack documentation” or whatever!
Def. to that, but I more meant your writing! Like: why isn’t this stuff we send out to all our users?!
Another stellar analysis. Very helpful and keeps me thinking about ways to stay ahead of "both" curves...