Marketing Is Stupid, Part 1:

Let’s Get Strategic

1_iNKQUI2vDCqZloA87T3mwQ.jpeg
What do I spend?
Where do I spend it?
Should I run Facebook ads?

These are the types of questions I get a lot — especially the first two. My response is usually the same: “Don’t focus on tactics. Tactics are the easy part. Make sure you’re focusing on the house rather than the tools.”

Another way of putting it:

Hole > Drill

This is something I’ve written about before.

If you are looking to build a house, you don’t seek out a “hammer guy”. You seek out the best builder you can, and reasonably assume that they know how to use the tools that are necessary.

If you are looking to build and scale a business, you shouldn’t seek out a “social guy,” or a “ppc guy”.

Business owners want a hole drilled. They don’t actually care what drill you use.

Marketing isn’t stupid by this definition:

1_ntThSiMK1IUS7JkVlEzEoA.png

However, it is stupid by this definition:

1_zC32GJLeJ_9XFivxiurS1Q.png

Actually, marketing is beautiful. It’s also behind the times. The two most innovative companies of the past century (Google and Facebook) are both, for all intents and purposes, ad tech companies. There are also a suite of killer marketing tech tools available. There’s a ton of backing to marketing tech startups and technologies.

But marketing is still archaic. Why?

Because it’s stupid.

This is the process I follow to make it a little bit less stupid. Typically, clients start with: “What do you think $X could get me?” Or, “Where do you think we should market?” (For the record, it urks me when marketing is used as a verb).

Them: This is like saying, “Hey, I want to make dinner. What ingredients should I use?”

Me: I don’t know, Jack (at the time of writing this, I don’t work with anyone named Jack). What dish do you want to make?

Them: Well, I think I can spend $30 on this meal. What could that get me?

Me: Let’s take a step back.

This type of thinking is akin to, “What should I spend?” “Where should I spend it?” “Should I do X?”

Here’s a lean startup methodology-inspired process I follow that solves this issue. Hope it helps:

1. Isolate Assumptions

When you start a business, or roll out a new product, you are inherently making assumptions about human nature.

If I start a coffee mug company today, I am implying that 1) I can do this better than every other coffee mug maker, and 2) I know I can do this, because I know what humans want in a coffee mug more than the other companies.

We all have the tendency to just roll with this thinking and start figuring out where we’ll sell these knock-your-socks-off-slap-your-grandma-and-call-her-skippy coffee mugs. But I find it helpful to take a step back.

We are hereby assuming:

  • These coffee mugs will be better than competitors. Why?
  • They will be better because we have killer designs. How are they killer?
  • Our coffee mugs are sick because our artists have been commissioned to make life-sized marionette doll replicas for heads of state (too weird?). Obviously they can make great coffee mugs.
  • We think we can sell a ton of these badboys.
  • Why do we think we can do that?

Write these out. Put em on sticky notes. Use a whiteboard. Whatever floats your sloop. We’re not getting tactical. We’re just surveying the landscape without making any more assumptions.

2. Define Hypotheses

(I’m writing a part 2 that will go out next week. We’ll get tactical in that post.)

Assumptions are things we accept as true without data. Hypotheses are testable.

Every initiative a business undergoes should fall in to 1 of 2 categories:

  1. Validating
  2. Scaling

When an initiative is in the hypothesis phase, you are looking to validate whether or not it’s true. In order to validate it, you need to define the appropriate metrics over a time period. I like 30 day windows.

Once the hypothesis is validated, you focus on scaling it. The validated stage is for minimum viable errthang.

Once you whittle it down to 2–3 hypotheses per category, figure out what metrics would give an accurate representation of the validity.

These are the umbrella hypotheses I feel everybody needs:

What’s your force multiplier/unfair advantage?

Everybody has one. Some have it more than others. What’s yours? This is not related to your team, drive, or ambition. It’s related to the problem you are solving.

Do people have this problem? (if yes, move on)

Is it a big enough problem that they’ll pay to fix it? (if yes, move on)

Do you have the right solution to the problem?

Don’t use surveys for this stage. People vote with dollars and their time. Surveys are kind of like giving a 16 year old their driver’s license based on their own assessment of themselves.

Metrics can vary to pretty huge extents for this phase.

Audience hypotheses

Who are you talking to?

Do have the problem that you identified?

For startups, this is usually pretty small data sets. Identify your hypotheses, then figure out where those people hang out. Then go meet them in person and pitch it to them. If they don’t dig it, they’re probably not the right audience.

Messaging hypotheses

Once you find out who you’re talking to, then figure out what you’re saying to them.

Does it resonate?

Does it move the needle?

Does it influence them?

This is isn’t the tactical post, but I’ll dip my toe in that water.

  • AdWords CTR tests

Pick 3 headlines, spend the minimum daily budget on Adwords, and put the headlines as the copy. Create a blank landing page that just says, “we’re under construction.” and have an email list signup form. It’s not important that you get emails. The only thing you care about here is the CTR (click through rate) on the ads. The highest CTR wins.

When Tim Ferriss was writing the 4 Hour Workweek (one of my favorite books), he couldn’t decide whether or not he wanted to call it “4 Hour Workweek” or “Selling Drugs For Fun and Profit”. He employed this tactic, and ultimately picked “4 Hour Workweek”.

My, how the world could have been different had the latter won the competition.

  • Landing page tests

Once you do that, then build a 1 page site using the winning headline, with 2 paragraphs of copy, 3 images, 1 video, and a call to action (form, quiz, ebook, whatever). This is where you test the other messaging elements: secondary messaging, taglines, narrative structure, etc.

You can manually create landing pages, use Google Optimize (it has a killer free version), Optimizely (no free version), Unbounce (DEFINITELY no free version), or anything of the like. Here, you’re testing call to actions. Whichever combination has the highest, wins.

Acquisition hypotheses

How are you actually going to get them to buy?

This includes everything from acquisition channels (Facebook, Google, Twitter, etc.), to offers (discounts, quizzes, ebooks), to some other stuff that we’ll delve into with the tactical post coming out next week.

Here, we want to test with a very limited budget. Rather than throwing $20k at it straight out of the gate, test it with $500. You already have the other elements. This should be the easier part. Here’s how you decide what to spend*:

*note, It’s not important that these estimated metrics are super precise. It’s just important to get the car moving so that you can steer it along the way. Don’t spend more than 5 minutes estimating.

  • A) Estimate lifetime value: How much money is that customer worth to you over their lifespan?
  • B) I like to target the cost per acquisition to be roughly 10% of the lifetime value. Do what makes sense for you, but decide on a realistic target. You can always reduce the cost per acquisition later.
  • C) How many customers do you need to acquire to be statistically significant? If you’re selling coffee mugs, and I target 10 people, and 2 buy, that doesn’t mean I have a 20% conversion rate. It means that my 2 friends wanted to help me out. Get outside of your friends circles.
  • D) Multiply C by B to get starting budget
  • E) Once these hypotheses are validated, pour gas on the fire.

4. Test

This is where you operationalize this. I’m gonna go into this more in the next post.

5. Scale And Iterate

If something isn’t validated, that doesn’t necessarily mean it is invalid. Don’t throw the baby out with the bathwater. You very well might be just a couple of tweaks away.