Highly visited websites can run multiple experiments and variations at the same time and see statistically significant results in a matter of days or weeks. This produces faster iterations, more tests, and more data proven results.
In contrast, sites that optimize conversion rates with low traffic can still run the same amount of tests and variations but they won’t produce the same statistical significance without adding copious weeks of time to the experiment.
With lengthy time requirements, would you even want to run experiments on low traffic sites?
This Insight will convince you that, yes, you can optimize your site and get bottom-line results without having a high traffic site.
Should I Focus on Boosting Traffic or Boosting Conversion Rates?
This is a classic chicken and egg situation. Which came first for successful sites: strong traffic or strong conversion rates?
So, where should you invest in your ecommerce site first – Search Engine Optimization (SEO) or Conversion Rate Optimization (CRO)?
Neither SEO nor CRO alone is going to magically take you from a low traffic, poorly converting site to the next Amazon.com overnight.
Start with figuring out your baseline traffic and ensuring you have a working site where you can track conversions. From there, expert CRO and SEO professionals can help take you to the next level no matter what your constraints are.
CRO and SEO have one thing in common: O for Optimization.
We recommend to clients participating in our proprietary CRO initiative, the Conversion Growth Program™, that they continue to run as much traffic-building and SEO programs as they can in parallel with our CRO efforts. We provide tools like a traffic and conversion calculator to estimate what even moderate changes to traffic and conversion rates will do for their revenue.CRO and SEO have one thing in common: O for Optimization. Click To Tweet
The Best SEO Can Often Be CRO
SEO experts at Moz recently wrote about the glory of combining SEO with great on-site UX experimentation and CRO. They argue that you will get better return on investment (ROI) on your SEO efforts if you’re driving that organic traffic to a site that has been through CRO. This is because the content on the site will be better aligned with the consumer and their behavior.
Conversion Rate Optimization can make sense for every site, no matter how high or low the traffic. If you manage a low traffic site and you’ve decided it makes sense for you to run some CRO programs, read on for our recommendations to get you started.
Statistical Significance? What Does That Mean?
Data analysis is still a red-hot skill in the business world because decision making can be backed by statistical evidence. But data analysis can be mystifying to the masses of business professionals. Terms like Baynesian probabilities, p-values, confidence intervals, and, of course, statistical significance only add confusion to simple concepts.
So allow me to demystify statistics for Conversion Rate Optimization (CRO) into simple terms:
We have the control version, or the original version of a page. Let’s call this Version A. We have a hypothesis that making a change to Version A will improve it – we’ll create this variation and call it Version B. We have conversion goals set up to track against this hypothesis.
We let the experiment run for a month and it produces the following results:
Version A: 100 goals reached / 5,000 visitors = 2% conversion rate
Version B: 250 goals reached / 5,000 visitors = 5% conversion rate
In this example, with Version B, there is a 150% improvement over Version A.
In order for data analysts and statisticians to prove that the conversion rate in their test is more than just a fluke, they need a sufficient number of visitors to participate in the test. This is the required sample size and it’s calculated by several factors including number of variants and the acceptable level of statistical significance. 90% statistical significance means we are at least 90% confident in our hypothesis (that B is greater than A).
In other words, we are at least 90% confident that version B will outperform version A by 150%.
The more test variations there are, the more visitors are needed in the sample size to reach a point of statistical significance. That’s why having more traffic produces statistically significant results faster.
Low Traffic Site Managers: Start Here
Sometimes you just don’t have the site traffic to get to statistical significance in a timely manner. You need immediate results and can’t afford to wait 28,134 weeks (542 years!) to reach conclusive results.
To address this traffic deficiency, here’s how Moz has recommended conducting CRO experiments:
- Make big changes to your page layouts
- Test goals with better conversion rates
- Reduce the number of variations
- Test the highest traffic pages on your site
- Test many things at the same time
- Test every page at the same time
- Start now and be patient
All fine ideas. With this approach, statistically, you will likely reach conclusive quantitative evidence. You will have data-driven decisions customized to your site and your audience.
Unfortunately, there are some drawbacks to Moz’s recommended approach:
- It will take patience to wait for results within a time period on par with a Deep Thought computer.
- If you’re a small company, you cannot afford to wait for these results.
- You also may not be able to afford development and implementation of big changes to your page layouts.
- Since smaller, incremental optimization areas are non-existent, your tests may be subject to false positives caused by complex multi-variant tests in an A/B split. For example, the cause of the winning test on one page may not have the same successful outcome on another page too.
CRO Roadmap for Low Traffic Sites
Start by ensuring conversion killing tactics are eliminated, and common roadblocks are removed. A free Stuck Score™ will identify where these common roadblocks appear on your site in a manner of minutes.
Embrace qualitative analyses with supporting anecdotal evidence. Trends and patterns can emerge that help identify optimization areas custom to your audience.
Qualitative analyses include session recordings, user testing, heuristic evaluations and any other method that produces valuable feedback without statistical, quantitative data. An example can include setting up a temporary experiment on your site and running a user test for feedback on the test. Listen and watch for feedback then iterate and repeat the test.
With a lower emphasis on quantitative results, you may need to implement experiments with engagement as your primary goal. The engagement metric tracks general interactions and activity, rather than measuring click/pageview goals for statistical significance.
Conversion rate optimization expertise can guide the optimization recommendations to implement these changes quickly on your ecommerce site. The results will show on the bottom line.
What’s the Minimum Amount of Traffic Needed to Work with a CRO Firm, Like The Good?
Any traffic is better than no traffic at all. The beauty of the Conversion Growth Program™ is that we can create custom roadmaps–approaching each client differently–to iterate our way to a ROI.
We’ve had lots of success in customizing our program for clients with lower traffic by focusing more efforts on qualitative research, bringing the consumer into the mix (through the list of methods mentioned above) to ensure the site is resonating with them–which can be done at any traffic size.
As a result, you’ll have a better ROI on your site when you start spending budget to drive traffic through more intense SEO and traffic generation efforts.
Get started now with a free Stuck Score™ to uncover the hidden roadblocks on your site.
About the Author
Dan Weinsoft is the former Director of Conversion and UX Strategy at The Good. Dan and the team at The Good made a practice of forming key strategies to boost online ecommerce and lead generation performance using testing, optimization, and data-backed insights.