If it's important, don't hack it

Growth hacking is a dirty world. Scraping websites and spamming activity feeds grows a business in the same way anorexia solves weight problems, swapping sustainable solutions with short term kludges.

The Faustian bargain of the internet is that you can swap your credibility for attention anytime you want. It’s not hard to get your name/product/company/brand well known, even if you’ve done nothing of note, but getting it well known for the right reasons is a different challenge.

The well known growth hacks used by sites such as SocialCam and Viddy require users to provide Facebook authentication to view a video. It works really well if your measure of success is Accounts Created. Here’s how well it works:

High and to the right, right? Who’s gonna argue with that?

Users will. Facebook will. Trying to hack social sharing pisses off the customers you trick, and the network you use. Lets see how that worked out for them:

This is the problem with false proxies. Someone in the company decides that accounts created is their metric of success, so the team works out ways to hack that number. Over time people lose sight of what’s actually important for the business.

Google+ growth-hacked their way to 170 million users, by putting interstitials in Gmail where the easiest way out of it was to click “Okay” which both set up an account and connected you with friends. Had Google+ focussed on a metric of “Users who choose to go to Google+ and share at least 2 updates per week“, they would have focussed their efforts on delivering value to users rather than popups that drive numbers.

Hacking a metric

Lets say you hire an analytics consultancy for your project management app. They’ll dig deep into your data and come out with something like the following:

Users who have invited 2+ teammates and posted 3 updates will upgrade to convert to paying users.

A Growth Hacker hears that and thinks “Sweet, all we need is a to get everyone doing this. Let’s go…”

What happens next? Users click the button, and the metrics go up, but the upgrades don’t follow. Lots of things correlate with users upgrading, but artificially triggering the correlation doesn’t cause anything. If it did, fires would spontaneously break out near groups of firemen.

What harm does it do?

Tricking your users just so you hit your metrics causes long term, if not permanent, damage. Bill Bernbach famously said that “Great advertising makes a bad product fail faster; it gets more people to know it’s bad.” This principle applies equally for web applications. Tricking potential customers into doing something prematurely, whether it’s creating an account, tweeting, or completing a task has the same effect. They’re likely to do it once, leave, and never look back.

Meaningful growth based on meaningful metrics

A meaningful metric captures a moment of value for the customer and the business. There should be no way that it can increase without both parties receiving value.

For Shopify, the total number of stores created is a bad metric as it can increase with no new value to the business. Alternatively, for a company blog things like pageviews or shares are equally bad as they can increase with no new value created. This is how the blogs evolve into junky list posts, cat pictures, news-jacking, and other crap. The metrics will go up but with no new visible new value to the company, and often some hard to quantify brand damage.

Shopify probably track something like stores with over $1000 in monthly revenue. That is a good metric as it only increases when proper customers are acquired, and barring fraud, increasing this figure is guaranteed to be good news for the company.

The danger of bad metrics can be seen in the old adage “what gets measured gets done”. Measuring irrelevant figures and your team start performing irrelevant tasks to increase a number that doesn’t really matter. Seth Godin said it best: “Gaming the system is never the goal. The goal is the goal.”