A quick repost of my guest post for Andrew Chen’s blog yesterday.
After recently moving on from adventures building a consumer gaming portal at Mochi Media (acquired last year for $80 MM), I’m now working on a new startup called Connected, which is a SaaS contact management solution for professionals. I decided to blog some of my thoughts based on my experience thus far with deciding on the right user acquisition channels to focus on.
When does ad buying work for SaaS businesses?
It’s a convenient belief that after you decide to build your software as a service (SaaS), Google AdWords and other networks will enable you to outsource all of your marketing efforts and focus less about user acquisition. This is not always true. Here’s a “napkin math” model to quantitatively decide whether or not ad buying is right for your startup based on reality, not guesswork.
A model for user acquisition
Paid user acquisition works for you when the following proves true
- LTV > CAC
The lifetime value (LTV) of your users should exceed the cost of acquisition (CAC) to get them in the door. As a reminder
- LTV = Expected Life x Average Revenue Per User (ARPU) x Gross Margin
In addition, for SaaS, you care quite a bit about costs and conversion rate for your funnel to trial, and from trial to paid. In specific, these look like
- CPC — cost per click to get traffic
- % trial conversion rate — users who convert to a trial of your product
- % paid conversion rate — users who convert to paid account
To estimate your cost of acquisition, you can base it off of estimates for your trial and paid conversion rates.
- CAC = CPC / (% trial x % paid)
An example of cost of acquisition
Let’s pick an example and work backwards. Let’s say you have a
- $20/monthly subscription
- 5% paid conversion rate — from trial to paid
- 10% trial conversion rate — from visits to trial
Then let’s pick a two different points for cost per click
- $0.50 CPC
- $2.50 CPC
In order to get a user at these CPC points
- CAC = CPC / (% trial x % paid)
- CAC = $0.50 / (10% x 5%) = $100
- CAC = $2.50 / (10% x 5%) = $500
In this example, it costs anywhere from $100 to $500 to get a single paying user at $20 per month. If you were trying to acquire 100 users ($2000/month), at $0.50 CPC that’s $10k ad spend, and at $2.50 it’s $50k. Drew Houston from Dropbox brought up very similar issues from his Dropbox Startup Lessons Learned presentation, where their initial search marketing test had a whopping $233–388 cost per acquisition for a $99 product!
Compare this against lifetime value
Compare this against the lifetime value of your user, or the total amount of profit you expect to receive over the user’s use of your product. This value should factor in the churn that you’re seeing from users canceling their subscription over time as well as what the payback period and working capital which you expect. Even though you might expect a user to be retained over a period of years, most startups don’t have the capital necessary to tie up their money for that long.
Let’s go back to the example above. We have the two users who cost
Assuming zero churn and zero operating costs on their $20/month subscription, you would recoup your cost on these user over a fixed period of time
- $100 / $20 = 5 months
- $500 / $20 = 25 months
In the case of second user, it would take over two years to recoup the initial $500 you spent to acquire them. You can offset this issue of working capital by setting the value at the amount of revenue you receive over a fixed period of time, or by being more aggressive with pushing them to prepay for longer periods of subscription cost upfront.
For example, what if you could get these users to pre-purchase their $20/mon subscription for $149/year? You’d be able to recoup the first user’s cost instantaneously, and get back a significant percentage of the second user’s acquisition cost.
Making the model work
The path to achieving profitability looks like making the model of having your cost of acquisition beneath your lifetime value work. You can quickly get a back of the envelope idea of whether paid acquisition is for you based on the examples and model above.
Doing this will help you determine whether or not you can profitably use ad buying as a source for getting users. You can also fine-tune your model to incorporate even more granularity such as
- traffic source
- working capital
Trying paid acquisition on for size
Now that we have the framework down, the question is whether or not paid acquisition works for you.
If this works for you, then congratulations- you are on the path to scalable riches! 😉 If it doesn’t work, then you should think about how far off it is. Getting ad arbitrage to work out profitably is extremely sensitive to changes in the steps of your conversion funnel, as well as the source of the traffic. So if you’re not many factors off, it may make sense to spend a few months refining your funnel and trying to optimize the channel the traffic is coming from. Here’s a few things to consider-
Does the math work?
Once you launch your product and get a sense of what the conversion rates are in each step along the funnel and the churn rate, it may be that the math doesn’t work out. If you’re not too far off, then it may be worth spending time trying to make the metrics work out through landing page optimization, increasing conversion along the steps of your funnel and trying to optimize your traffic sources. However, if you’re several factors off (this is common in highly competitive markets) paid acquisition may not make sense as a strategy for you.
Is your product in an existing market or a new market?
Intent-based paid acquisition channels like search advertising work best in an environment where users are aware of the problem and actively searching for solutions which your product meets. You can look up potential search terms and volumes through Google AdWords Traffic Estimator, including estimated average cost-per-click and monthly search volumes. If not, you can also experiment with targeting sites that reach the demographics of your users.
How much working capital do you have?
While theoretically you might be willing to pay up to the full LTV of the user, you may want to limit the amount you’re willing to pay based on a fixed time period, for example the expected value from the user over 6 months. This may be because at some point you run into working capital issues paying for users who may take years to break even.