-
Website
http://andrewchenblog.com/ -
Original page
http://andrewchenblog.com/2009/01/19/how-to-create-a-profitable-freemium-startup-spreadsheet-model-included/ -
Subscribe
All Comments -
Community
-
-
Top Commenters
-
Popular Threads
One thing that we've found very useful at my startup HubSpot is looking at product management and feature prioritization from the lens of the three primary variables in the equation:
1) Does it reduce cost to acquire customers?
2) Does it increase the life-time value (by raising ARPU or reducing churn)?
3) Does it decrease COGS (cost of goods sold)
What we've found is that at various stages in the business, different things should be the focus.
1) acquisition
2) engagement
3) retention
4) monetization
as 4 different areas, although as you say, 2, 3, and 4 are the key components of LTV.
The other macro-issue to balance is market strategy - because the value of a business is both revenue but also valuation multiples! So you want to make sure you're maneuvering the business to a market sweet spot.
Bob
------------
Bob Thomson
storm ideas
http://blog.stormideas.com
http://www.colaab.com
twitter: movingforwards
It rivals what others charge a lot for.
Thanks!
Likewise, ignoring cash flows is probably the single biggest reason that businesses fail. Ultimately, cash is king.
The biggest issue with the model that I found is that it doesn't account for the large upfront fixed cost of the game, the infrastructure, etc. That's a huge burden on the LTV of the customer and the bottom line. Even a 'cheap' game made with a 500k investment would make it almost impossible to achieve a positive result with the numbers you give
BTW, I enjoyed the article and appreciate the model, but it's a mistake to dismiss the complexities of investment and cash flow when making a decision about your business model.
Appreciate your honesty,generosity and detail in offering this up.
Look forward to reading more on your blog.
Appreciate your honesty,generosity and detail in offering this up.
Look forward to reading more on your blog.
Scott Kilmartin
- - - - - - - - - - - -
Designer / GM
haul
http://www.haul.com.au
http://www.riveting.com.au
http://www.twitter.com/scottkilmartin
The learnings from those areas are definetely helping me with my current startup.
http://spreadsheets.google.com/ccc?key=pQWEn2YA...
Obviously see complementing Dave Mcclure's AARRR presentations(how about connecting these two , and similar thoughts for 1 comprehensive "whitepaper" ?)
BTW, you need to look in detail at EACH traffic generator (leads, ads, viral, press etc) to learn what and who works better and focus on it. Lots of work, but worth every penny (or cent).
We have a website that has decent traffic (Alexa rank #407,000), all generated by content, not by advertising.
We do have people signing up for the free trial, but much less than we expected and what you suggest with your 20% and 10%.
Thus something is missing or should be improved.
1) The model uses "time-periods". Do you prefer to do your calculations monthly/weekly, etc?
2) LTV is very sensitive to retention-rate, and the model "defaults" to 80% retention period-to-period for all types of users, which seems incredibly high, even if your time-period is weeks. (At least across ALL users, as opposed to the highly engaged)
I realize it's meant to be tweaked, but I'm looking for a little extra insight here. Thanks!
For #2, the retention rates completely depend on your product - 80% isn't high for installed software or the stickiest websites, but something like 20-30 might be more appropriate for the basic consumer internet site. Just depends.
Ah... for #1, I've got my myopic startup blinders on apparently.
W/ respect to number 2, I'm remembering back to your presentation w/ Daniel James, where (unoptimized) Whirled had a 35% week-to-week retention rate across all users. Very hard to ferret out other examples of week-to-week numbers (that could be used to make the model work, hopefully successfully) so I was curious if the 80% had any 'meaning' to it, but it sounds like that was just my social-gaming blinders on again.
So just because the Whirled numbers were 35% doesn't imply anything about the month-to-month, since it depends on what grouping or subgrouping of users come back. Now they are probably correlated in reality (the above situation is probably a corner case) but still worth thinking about.
Thoughts?
(Great work BTW Andrew!!!)
I want to put some emphasis on the importance of ease-of-use for customers of a freemium based service.
Knocking even a few seconds off the time it takes a customer to go from free to paying for something can *dramatically* increase paying customers.
A great example is south Korea. The virtual currency market here is very well developed, and it is extremely easy to get money from a bank account into a website. Sometimes it takes as little as a single click and two lines of information.(!)
One nitpicky edit: On the "retention" tab, under the "total users" matrix, you accidentally labeled what should have been "cumulative total users" "paying users."
really thank you, that was great.
I've blogged about the process and shared the spreadsheet here: http://dubitplatform.com/blog/2009/8/31/templat...
I wanted to say a quick thanks for sharing your insights, and thought you'd be interested to see how they've been used!
Matt
https://www.nppa.org/professional_development/b...
I like the article, and love the general usefulness of the spreadsheets.
But, I agree with Steve that a constant retention rate seems a bit off. I am thinking that most sites would have a decay model following a logarithmic formula, and looking at the model you used, the 20 time periods sort of implies a 20% monthly loss, which winds up giving you a less than 10% retention at the end of a year. Else you're analyzing over 20 years. All of which confuses me.
Do you have any suggestions for how to come up with a logarithmic formula to project customer loss from a known retention rate?
Kirk