Stolen directly from this lecture series, I’ve updated my mental model on when to use ads. Instead of having you watch many 40+ minute lectures, I’ll distill it down into a few paragraphs below.

Old Model

  1. Get customers.
  2. Get up to $1K/mo so I can afford ads.
  3. Pay for ads.

This is an incomplete model because it disregards a few things:

  • Customers churning. $1K/mo is good, except it’s $0/mo next month if my churn’s 100% and I don’t acquire new customers.
  • 111 x $9/mo customers has a different risk profile than 10 x $90/mo customers. $90/mo customers churn less generally speaking but you only have 10 of them. So any cancellations injects more volatility to your MRR.

Debugging My Startup / Products

Here’s my hypothesis of symptoms vs problem to be solved. Presented in format “symptom = problem to be solved” below:

  • No new customers = Landing page sucks, or no product-market fit (PMF). Work on sales / marketing.
  • “Gimme free trial” = Targeting wrong / generic customer base, or no PMF. Work on defining your avatar more closely and finding channels to market to them more directly.
  • High churn (varies depending your model, but I reckon >10% is ‘high churn’) = Product sucks. Work on talking to customers and improving your product (ideally paying customers).

New Model

If your churn is below 0, you should pay for ads.

WTF, how can you have below 0 churn?

This is when your early adopters are recommending more folks to use your service more than the amount of people that are unsubscribing.

When you can do this, you intuitively have a growing business model (more paying customers gets you more paying customers).

With a churn >0, you need to actively acquire new customers which takes away from your ability to improve your product. This is bad because large competitors will always have more resources to improve your product faster than you. So you’ll lose. A churn >0 is almost by definition a dying business model since doing nothing will mean your business going to $0 MRR.

Thus, the main objective here 2 twofold:

You can make your product very good compared to competitors, so people will feel stupid using ANYTHING else. Achieved in a few ways:

  1. Bottom-dollar pricing (not recommended).
  2. More value given the price. More relevant features / each feature is more refined / etc.
  3. Killer and unique offers. Something like Jasper AI might offer 50 kindergarten-wrappers around GPT3 / ChatGPT but it addresses a wide breadth of copywriting needs and long-form writing. Wraith Scribe specializes in long-form writing and has semantic keyword suggestions + AI-powered editor + article length picking and the feature set goes much more into “how to create a a high-quality article as quickly as computationally possible, and how to allow users to edit and perfect their article as quickly as humanly possible.”
    1. If your focus is blog / long-form writing you are simply stupid to use a non-specialized tool like Jasper. If your focus is more short-form copywriting, then Jasper is great.

Caveat

This is ONLY for software startups. Other products, like e-commerce will have different models of when you should use paid ads. Though I’d probably use Tik Tok or something to gauge demand before wasting money on ads to be honest.