Most startups these days use some form of analytics (and if they don't, they should.)
They install a little tracking snippet on each page of their site or app. They add in various events, and they start to get a visualization of what is happening as it’s happening.
They then assume that these analytics can serve as growth projections.
This is a mistake.
Event-based analytics are historic. They tell you what has ALREADY happened.
This is pretty cool, but companies don’t succeed because they can tell others what happened in the past.
They succeed because they plan effectively with the future in mind.
Throughout this entire section, we’ve been analyzing all the horrible things that can happen to your growth over time.
From feeling the churn to hitting your carrying capacity, things can get ugly fast if you aren’t prepared.
The point is not to scare you into giving up viral marketing altogether, it’s to give you the lay of the land so you can better navigate it.
With these insights, you are ready to look into the future and predict what may come, then expertly adapt to ensure nothing holds back your growth.
The Past Is The Past
Leading indicators are metrics for events that have happened and act as predictors of lagging indicators (other metrics that are difficult to predict otherwise).
By using a leading indicator like customer complaints, a good data-driven business can infer what their churn rate will be in the near future.
If complaints are up, your churn rate soon will be as well.
If they’re down, an improvement is likely in store.
However, even using leading indicators, most growth projections are just wishful thinking in disguise.
They’re often used to hype up new employees, investors, or partners – but have no basis in reality.
Even when done with sound data science, predictive analytics aren’t 100% accurate – but they can be incredibly helpful nonetheless.
Famed statistician George E.P. Box once said, “All models are wrong . . . but some are useful.”
This has proven true time and again.
Modeling User Acquisition Over Time
After learning a bit about viral marketing, growth, churn, and carrying capacity, the logical next step is to tie them all together to help you derive some business value from what you’ve ascertained.
The goal here isn’t to give you a 100% accurate prediction.
It’s to give you something useful that may be able to help you avoid the disaster of hitting your carrying capacity.
To get started, let’s set up a scenario:
- i = 4
- conv% = 10%
- Given these two, K = 0.4
- ct = 1.0
- t = 20.0
- u(0) = 10,000
Next, let’s introduce a new concept into the mix – non-viral marketing.
Let’s name our new variable nm, and use nm(t) to signify the amount of users our non-viral marketing efforts will yield within a specified number of days.
Let’s say our non-viral marketing efforts yield 200 users per day for our new app. Since our t value is already set as 20.0, we’ll use that here:
- nm(20) = 200 * 20 = 4,000
Now let’s try to predict growth over time using the following equation:
u(t) = {u(0) * [K^(t/ct + 1) – 1] / (K – 1)} + nm(t)
This basically takes our equation for projecting viral growth over time and adds the number of non-viral users acquired.
u(t) = {10,000 * [0.4^(20/1.0 + 1) – 1] / (0.4 – 1)} + 4,000
u(t) = {10,000 * [0.4^(21) – 1] / (0.4 – 1)} + 4,000
u(t) = 20,667
Okay – so using the above variables, after 20 days, we should have right around 20,667 users.
Given that we gained approximately 4,000 users from our non-viral marketing efforts, this means we acquired 6,667 users from our viral loops.
Not bad, right?
Well, it wouldn’t be bad at all . . . IF the projection equation were correct.
Can you spot the issue?
Fine Tuning Our Equation
The issue is that with most viral engines, whenever a non-viral user is acquired, they seed your viral loops similarly to a user who joined via a virally-acquired user.
Right now, this equation simply bolts them onto the end.
In fact, each one of these users will result in virality themselves.
To correct this, your equation will change in two ways:
- You can ONLY run this equation when t = ct
- At the end of each pre-defined ct interval, you will need to rerun the equation again while updating your u(0) value to your previous u(t) value.
So let’s try this again. For simplicity’s sake, let’s keep ct = 1.0. (Remember, nm(1) = 200.)
u(t) = {10,000 * [0.4^(1.0/1.0 + 1) – 1] / (0.4 – 1)} + 200
u(1) = 14,200
After a single day, we’ve grown from 10,000 to 14,200 users.
Wow! That’s incredible!
But don’t throw a viral party yet, because this technically isn’t growth.
It’s user acquisition.
The Difference Between Growth and User Acquisition
Let me clarify something most amateurs masquerading as experts and thought-leaders don’t tell you.
User acquisition does NOT equal growth.
User acquisition is only a piece of the overall pie chart of growth.
Once we acquire users, we still need to keep them.
If we don’t keep more, then we lose (and we WILL lose some), and we won’t grow.
You WILL lose a percentage of the users you’ve acquired every single day.
This is called your churn rate.
If you’re not carefully tracking your churn rate in comparison to your growth rate, you’ll risk hitting what’s called your carrying capacity – or the point at which your growth rate equals your churn rate.
In other words, your carrying capacity is the mathematical moment at which your growth becomes flat.
In Summary
Most novice – and even seasoned – founders and growth engineers will simply assume they can always find a way to acquire more users faster to stave off their carrying capacity.
As such, they believe their carrying capacity is unlimited, given enough money and resources.
This couldn’t be further from the truth.
In fact, we just talked about the creeping surprise of network saturation, new competition, and the various other factors that may affect your ability to acquire new users in your existing channels at the rate you’re used to.
The most prudent thing to do is to place your best people and your most powerful resources on continually retaining users and reducing churn.
Then and only then is continued growth more of a sure thing.
What’s Next
Now that we’ve covered some of the methods you can use to sidestep falling into stagnation or, worse, reverse growth, we can head in a more positive direction.
A really, really positive direction.
After all, the goal of any good growth engineer is to hook users. How do we do that?
By getting them addicted to your product, of course.
How Do You Get Users Addicted To Your Product?
We’ve talked about viral hook points in the past, but now we’re going bring it to the forefront and engineer that sweet spot when user action meets user desire.
All to the benefit of your viral engine. Get ready to go fishing!
TEACH ME HOW
What did you think of this article?
- Need any help understanding our math above?
- What are some of the ways you can think of to reduce churn?
- If Viral Panda had an arch-nemesis, who would it be?
