Why Split Test?

by Brian Toomey, JB Analytics CEO

At the top line, we test because It makes more money for your business. Period. 

We are all wrong more often than we think. Especially people that don’t think they are wrong –  are more often than they think. 

Designer vs Data

The smartest designers and marketers are still less smart than user data with valid testing. Science – i.e. iterative testing – makes us smarter and better at design in the future because we know more about our users and what they like.

Your Bottom Line

  • Testing makes you money. Often it’s best to speak with a business/marketing team to determine a value for a given KPI (i.e. a sale, enrollment) phrased in terms of a willingness to pay or change to bottom line revenue, i.e. ‘each submission is worth about $100 to us.’ This allows us to convert conversion rates (like 1% of users filling in a contact form) into expected marginal monthly or yearly revenue at current traffic levels.
  • The smartest companies are doing it. Amazon and Google are continually and deeply testing.
JB Analytics has been a great help to charity: water. They combine a passion for data with an equally intense passion for results in the digital space. We recommend them to anyone with business intelligence, analytics or SEM needs.

Small and Steady

  • It allows iterative, continuous improvement. Speed of business improvement is directly proportional to the speed and quality of your experimentation. 
  • In most cases, it’s better to test many small changes quickly, rather than one huge version change that takes months to build and then doesn’t do as well as hoped. This also lets you see the individual, marginal contribution of each change rather than letting ten positives and ten negatives wash each other out. That said, some times bolder changes to break out of a rut or maximize information gleaned from relatively small traffic volumes are needed. 
  • Mistakes can be just as valuable as learnings (but only if you test with a clear hypothesis, track results carefully, and accept when you are wrong). Every test teaches you something about your users.

Other Considerations

  • Disputes can be settled without animosity or office politics. Rather than choose one opinion over another, you can be guided by real data. Decisions are in the hands of your community of users, voting with their actions. 
  • On the internet, so much is measurable. Your organization can benefit from the same precise data science used in modern clinical trials. 
  • Testing is incredibly cheap! It’s almost always much faster and cheaper to lift conversion rate by 30% than it is to acquire a steady stream of 30% more users at the current quality.

Happier Users & More Traffic

We’re passionate about delivering value through design, data and development. Let’s talk!