Failure comes from chasing success.

The best thing that ever happen to advertising has been the arrival of data and accountability, but it’s also been the worst thing. For all of us who grew up on advertising before digital, knowing that results were vague, that success happened over several years and that brand metrics would survive a mistake wasn’t that comforting.

But it did allow us to focus on our gut and find ways to connect with people using creativity and empathy. We all loved the objective new world of digital, a chance to see what worked, to optimize, to see the difference we make, but what if results and measurability have made us fail?

The devastating uncomfortable truth of advertising is that we have no idea what real-life metrics are shifted by what we do. There are simply far too many variables that we can’t control. We once embraced softer metrics, like awareness and favorability, but these take so long to measure and change. They are so hard to attribute, we’ve fallen in love with the instant gratification and objectively of digital metrics, it’s a world of clicks, views, “likes,” visits, and such. This is where part one of the McNamara fallacy kicks in:!

1) Affect the statistics we measure

We may set out with a brief to build awareness, but when all we can measure are views or “likes,” that’s what we seek to accomplish. We unknowingly move from a broad strategy into tight tactics, tactics that may accomplish your goals, but little else.

The metrics of measurement suggest that the toilet cleaner driving “likes” by publishing cat or baby pictures is doing well, yet common sense would show this to accomplish little. The Internet is now awash with beer brands showing they drive conversation by asking people how their week is going, by brands offering prizes for a retweet, it’s a mindless, vapid world of selfie competitions, drone vitals and one-off vending machines. The lofty goal degenerated into proxies for demonstrating success that at best are meaningless but at worst are damaging.

2) Fake’ vertising

Companies that work with influencers to ensure that viral videos are published by the connected trendsetting people that matter have morphed into companies that guarantee views. We’ve swapped quality for quantity, and we’ve a tranche of companies moving more towards click farms than PR. In 2014, we can now buy 1m views on YouTube for $400, or 100,000 Facebook likes or 70,000 twitter followers, “success” has never been cheaper.

3) Convenient assumptions

The metrics we count have no meaning; we’ve no idea what a like is worth, how many people a retweet reaches, or what the value of that would be. We spend time and energy obsessing over precise data that has no understood value. Some studies show the value of like be as much as $214.81; others suggest it’s as little as $0. We get confused between the cause and the effect, Facebook would have us believe that the act of “Liking” Coke on Facebook makes them buy more Coke, because “likers” spend more. Common sense would tell me that people who drink more Coke, like Coke more and are more likely to like a Facebook page.!

Starbucks opened a tweet a coffee campaign, which produced $180,000 of sales. But we’ve no idea if these were incremental sales or just a way to rechannel existing sales through a overly complex, less profitable payment system.

There are case studies online that attribute social-media sites for being entirely responsible for selling incremental cars. Does this smell right to you?

4) Don’t care about things that do matter.

It’s part Two of the McNamara fallacy that really hurts, not only do you do what you can to boost what you can measure, but you fail to do anything that produces results you can’t measure.

We live in a world where the biggest changes come from the things that are hardest to measure, it’s the word-of-mouth recommendation after the good experience. It’s the smile at the checkout. It’s the TV ad that made you cry. It’s the Yelp review. I’s the new look to the store, the rebrand, the good press on TV, the new CSR program. Buying things takes a billion things to come together and align. Many of the most important things don’t happen online, they can’t be scene and they should never be forgotten.

5. Getting comfortable with ambiguity.

The relentless search for data and effectiveness prioritizes what is easily measurable. The goal for results ends up being the strategy. It’s a horrible conclusion, it sits funny and it looks terrible in a deck. But the goal for most organizations should be to use data to track progress and to diagnose, not to reward. Rather like education, tests are not there to teach, they are there to monitor just some, not always important, metrics that can designate progress or issues.

Or as William Bruce Cameron said: “Not everything that counts can be counted.”