Agreeing a client Life Time Value (LTV) framework

Prior, related post
Thank you to Owls @LemonadeAlpha @gregdocter @MrMadila @verto0912 for comments and questions on the previous, related post and supporting model.

Recapping background:

  • It’s important to agree a sound client LTV to appraise various growth initiatives against
  • Another way, we need to ensure our Cost Of Acquisition (CAC) is lower than our LTV over time to be profitable
  • The pior post and this post seek to create debate and discussion amongst Owls on this important topic, to hopefully reach alignment on a framework to use going forwards
  • This framework should not be set in stone forever and can be revisited at a future date, once we have more data from growth experiments and product traction
  • Hopefully the poll at bottom of the post will yield a clear, decisive vote and be a signal this proposed LTV framework is acceptable to fellow Owls. If this result is not achieve, more debate and discussion will add value and still move things forward

Proposed LTV framework
Based on the prior work on the model (Tab 2: “V1 ROAS MODEL”), followed by welcome feedback from the Owls mentioned above in the prior post, I propose we use these inputs in our initial LTV model:

  1. Average DPI units bought by client: 2.5 (as per data from @jdcook and discussed in first post)
  2. Average client holding period: 3 years
  3. Rate client is cross sold and buys other Coop products: 35%
  4. Average extra Coop products bought by client over 3 year holding period: 2
  5. Yield achieved via intrinsic productivity use cases: 1%

Hopefully 2-5 are conservative and we should run the numbers later in the year when we have more products live.

Example of LTV framework in action
If we use this LTV framework within a simple media spend scenario, we generate the implied results shown in Tab 1 (“V2 ROAS MODEL”) of the model and in this screenshot below. Note LTV framework components in green highlighted rows.

Screen Shot 2021-03-14 at 21.17.22

Next steps

  1. Please either provide more critique and comments below
  2. Or, if you’re ready to vote in the poll, please vote FOR or AGAINST the proposal

This is an important thing to use in our day-to-day growth work, should be debated and scrutinized and be robust, so all new feedback is welcomed. Thank you.

Vote on proposed LTV framework
  • FOR proposed LTV framework
  • AGAINST proposed LTV framework

0 voters


Hey @DevOnDeFi, thanks for keeping at it man! :+1:

Stupid question - what is this for? like why are we voting on this / do we need to vote on this?

In your original post, you said that we should “use this in our ad, content, referral, social, and other channels.” Basically, every time we are paying to acquire new users, this framework should be used. Most, if not all, of these things sound like they sit within GWG’s domain.

So, from my perspective, if heads of GWG + folks like you working on this stuff agree that this is the appropriate framework, then you should go ahead and use it. No vote needed. If you want more community input on some of the assumptions, perhaps that could be done at the weekly meeting or the growth meeting?

Does this make sense or am I missing something important here?


I fully agree with @verto0912 comments here. No vote needed. This is a model we need to justify ad spend in the short term. It is a in line with Index Coops principles of being data driven. Hence should be implemented. Additionally, this is an example of the importance of the AWG group without their input this would not have been possible. This model will streamline growth activities and make GWG activities more effective. I cannot wait to use this as a tool for success in the Index Coop. This model also serves as inspiration and a standard for the type of work I will be producing in the future for Index Coop.

1 Like

Hey @DevOnDeFi this is solid work. I’m glad that we are starting to shift our thinking as a Coop to become much more strategic. Seeing the breakdown of how DPI buys translates into revenue is very helpful for us to visualize what we are really getting from our spending.

I like that these forum posts can help serve as a repository for some of these approaches. A few asks:

  1. Can we build out that model more to include analysis of the following:
  • Exchange listing broken into small, medium, and large exchanges

*Institutional sized sales of $DPI ($1 million to $5 million DPI bought)

  • I’m also curious what this means from a LM framework. @Matthew_Graham how do you think this would best fit into the framework we have now?
  1. In my mind the next step once we use this model to build out analysis of our main revenue streams and sales opportunities the next step is to start building this analysis into our GWG fund prioritizations to help us ensure that we are spending money on the most LTVROAS activities (I dont know if that acronym is going to catch on :joy:)
1 Like

Great clarifying question, @verto0912.

If we need to agree narrowly in GWG, with leads, etc, that’s even easier for me. Given this impacts other things, which might not all fall into GWG in time, I thought worth reaching out more widely. A lot of that depends on what the GWG owns over the long term though. However, if there’s a steer to get at this in the GWG, that’s cool.


Thanks for the endorsement @Mringz ! I’m humbled.

Great question re extending this model to apply to other activities @BigSky7 - we can certainly do that. I’ll DM you re model exchange flows.

In time we can def. use this framework to appraise our GWG activities against. That said, some lower return items have a key role too - like different positions on a soccer field. Look forward to seeing the GWG do more of this kinda stuff.

Thanks for the comments and votes folks!

We’ll use the framework in the growth trenches and share any insights as they come.

1 Like

@DevOnDeFi Some separate thoughts:

  1. I think it makes sense to remove Intrinsic Productivity (IP) yield from the LTV estimate because that is intended to be passed back to holders, not captured by the Coop.

  2. I think the FLI (trading fees) suite should be modeled separately from the index products (streaming fees).

  3. Given how little data we have to construct LTV estimates at this point, do we have comparables for what Vanguard, Fidelity, Cash App, Robinhood, etc. have as LTV for a customer/user?

1 Like