You’re Not Failing Enough

I was asked to give a talk about paid traffic sources at 500 Startups last week. The presentation is embedded below.

Although I couldn’t resist diving into deep, specific, tactical stuff near the end, the three most important points I wanted to impress upon my audience were:

  1. Most online ad campaigns (even those created by professionals) fail
  2. The only foolproof way to succeed is to try (and fail) enough to exhaust every other option except the successful one
  3. Therefore, your objective should be to fail as quickly and cheaply as possible

It’s become common knowledge among the lean startup movement that you should launch quickly, iterate, pivot, etc. But I want to take this one step further as applied to traffic (and startups as a whole):

When you launch a campaign, your objective should be to make it fail.

When you launch an advertising experiment, it will most likely fail. The null hypothesis is that it fails. This is a good thing, because it creates defensible barriers to entry for your business.In other words, once you have a successful campaign, a novice with a $100 AdWords coupon won’t be able to disrupt your acquisition channels.

If chances are that your campaign fails, you might as well do it quickly and painlessly.

I know it seems crazy to set a goal of losing money. But just give it a try. Because here’s what happens when a campaign fails:

  1. The campaign failed because it spent money without bringing in enough conversions or revenue to pay for itself.
  2. If the campaign is spending money, it’s generating traffic.
  3. If the campaign is generating traffic, it’s also generating data: click costs, conversion rates, ad copy and landing page split test results, etc.

And as any good marketer will tell you, data is everything. He who has the most data wins.

Don’t aim for launching a campaign that’s instantly successful/viral/profitable. That’s a fool’s errand, and it can only lead to disappointment.

Your only objective with a new campaign should be to collect enough data to validate or disprove your assumptions.

Then go back to the drawing board, use what you’ve learned to create a new campaign that fails slightly less than the last one, and try again.

Don’t worry about the conversion rate or CPC with a new campaign. Just get the data, so you have a baseline you can optimize from.

If you get an additional data point about what works and what doesn’t you win, no matter the result.

Pickup artists call this mindset outcome independence, defined as “The mindset of not focusing on a specific result, or growing attached to any outcome.”

If you’re not attached to the outcome of a split test, you’ll never get demoralized by its inevitable failure. And you’ll never risk giving up on a traffic source or acquisition strategy too quickly because your first few campaigns failed.

This can be an incredibly powerful mindset. Embrace failure. Never stop testing. And the successes will come in time.

The high rate of failure for most ad campaigns is the reason we started MixRank. We built our startup to catalog and analyze millions of split tests and campaigns  so you can learn from your competitors’ mistakes rather than making them all over again.

I don’t post that frequently, so your best bet to get notified about new posts like this one is to subscribe by RSS.

  • Anonymous

    Technology is making it easier and easier for developers to create startups and disrupt long-standing business models, but the flip-side of this is that there’s more and more competition for scarcer attention making standing out very difficult. One minor point of disagreement: I don’t think that your objective should be to flat out fail because even when I didn’t know it yet, there have been times that I figured out the best form of advertising for a particular niche product right off of the bat by accident. However, I do agree that gathering data and weighing the costs and return on investment should be key. Each business reacts differently to different types of marketing and collecting data and analyzing it puts you ahead of the vast majority of companies. I know too many companies that just try and use one type of Adwords ad and thats it. Seriously, or they just try one of the companies listed at and then go "That’s it, we’ve maximized social media output." If you’re not experimenting and trying out every possibility you’re just wasting a lot of opportunities. It doesn’t take much effort to (at the very least) just use a simple spreadsheet to chart how much many dollars generates how many clicks and how many conversions. I think there are a lot of companies that are approaching the same problem that MixRank is in slightly different ways and this is a pretty interesting time to be focusing in on this type of analysis. The good thing is that all of this competition in the marketplace is inspiring companies to analyze their data more and more and better and better options for analyzing all of this data are emerging making this quite an exciting time for this industry.

  • vbelfor

    Staying objective and cool-headed and using all experiences as data points is a generally good approach to many things in life. Troubles start when one either has unrealistic expectations or begins to believe one’s own bullshit. As long as a person is willing to test her own beliefs, things usually turn out OK. One of my favorite sayings is "don’t believe everything you think".Good post.V –

  • Dentist Bountiful

    well, I guess this presentation paid traffic is worth the time because we will be learning a lot from it, right?