Conversion Optimization for Low-Traffic Websites
“Conversion optimization isn't possible for businesses unless they have millions of monthly visitors”
If you think that, stop.
Even if your website generates 200 sessions daily, you can still optimize it and get more revenue from the same traffic.
This article will…
- Show you how to optimize your site’s conversion rate with low traffic
- Explain three important A/B testing facts
- Describe five useful research techniques
- Point out where to look for speed and compatibility issues
- Provide testing and tool recommendations
A/B tests are not required to optimize for conversion
There is a common misconception around A/B testing and conversion optimization.
Some people think that if they run split tests, they are doing conversion optimization. The others believe that if they can’t do split testing effectively (because their traffic is too low), they are not able to optimize their website at all.
Both groups are wrong.
A/B testing does not equal conversion optimization. It’s a vital part of the process, but not necessary in order to increase your revenue (which is the ultimate goal of CRO).
If your traffic is low, you can just implement updates for 100% of the visitors straight away and check the results after 2-4 weeks (that should give you enough data and cover at least one full business cycle).
The action plan is almost equal to any conversion optimization process - the only difference is that you compare two time periods instead of running a split test. The same setup and follow-up - things like setting up data collection, research, wireframing, design, coding, setting up tracking and analyzing the results - are still required.
Simply compare the timeframes directly before and after updating the website. This is crucial to minimize the impact of external factors skewing the results.
Things to be aware of and account/look out for:
- Different weekdays (always compare full weeks, i.e. 7, 14, 21 day-periods etc.)
- Black Friday or Cyber Monday
- Seasonality (for example, most bikes are sold in spring. Most gym memberships start in January, etc)
- Special event for the industry (for example, the release of a high-profile video game)
- Major changes in your traffic generation structure
If you keep the external factors at bay when checking the performance of website improvements, your findings are likely be correct.
Running A/B tests with low traffic
Even with small traffic it is possible to run split tests.
There are three things to keep in mind when utilizing A/B testing for a low-traffic site.
1. You don’t necessarily need a fixed number of visitors to test
You might have heard that you cannot A/B test unless you have a certain (high) number of visitors. It doesn’t work like that - the sample size you need is a result of your current conversion rate, expected growth and the statistical significance level you are aiming for.
Use this sample size calculator from Evan Miller to see how big a conversion rate uplift you need to run one split test per month with your current traffic.
Example: You want to run a test with two variations. Your page generates 3,500 visits monthly, conversion rate = 4%. According to the calculator, you will need ~50% conversion rate uplift to reach 95% statistical significance in one month.
If reaching 95% confidence would result in running a test for 2+ months, consider aiming for lower value. 75% would be fine - it means that 3 out of 4 test results will be valid. This way you will end tests faster, check more hypotheses and eventually make more money (if you test the right things).
This only works to a certain extent, of course. Limit using this exception when you see the conversion rate increase is at least at 30% level. Also, some tools might show you 75% confidence when you only have a dozen of conversions per variation - don’t trust these either. Wait for at least one full business cycle (2 weeks) before jumping to any conclusions.
There is a bit of risk involved, but the ROI from accepting lower statistical significance level will be better than striving for 95% all the time.
“Big money tests” with high impact allow you to come to a conclusion more quickly and more confidently, even without a million visitors.
2. Using micro-conversions speeds up testing
You might not have enough leads or transactions to be able to run split tests, but those micro-conversions which actually lead to the big goals can be used as a success metric for an A/B test as well.
For example, split test your homepage by setting the click-through rate to the pricing page as a conversion. If this CTR is 18% while your overall conversion rate is 3%, your baseline conversion rate used to calculate test results is 6x higher. It will take way less time to draw a conclusion.
Important: Remember to check whether your revenue grows as well. The more visitors you bring to the pricing page, the lower their average motivation to buy the product might be. Increasing the CTR by 20% will probably not give you a 20% increase in revenue, but rather 5-10%.
3. Testing bold ideas is necessary when your traffic is low
Finding the perfect shade of yellow for the “Add to cart” button on Amazon might result in a 1% conversion rate uplift and eventually lead to a revenue boost, but if your small ecommerce is waiting for those results to come through, you’ll be waiting a while.
Don’t bother with small tweaks. Focus on the big conversion problems and test different solutions to find the one that brings the best results.
Important: if your research proves that there is an issue, don’t give up right after testing just one solution (even if it brings uplift). The way you tackle the problem can always be better. Squeeze out as much as you can from the conversion issue you discovered.
Check out “5 "Big Money" A/B Tests that will Actually Touch your Bottom Line” for more.
Mastering conversion research without much data
Low traffic means you will have to wait months until you collect significant amounts of data in your analytics tool.
Fortunately it’s not the only way of gathering insights and coming up with hypotheses.
Say hello to qualitative research..
To analyze a low-traffic website from a conversion standpoint, use:
- The 5-seconds test
- First click test
- Usability testing
- Heuristic analysis
- User session replays
The 5-second test
Optimizing for clarity and a good first impression is a vital part of CRO. Visitors need to know in the blink of an eye if:
It’s the website they hoped to get to
The content of the page matches what they saw in the ad/traffic source
It’s clear what the company does
To check clarity, use the 5 second test.
UsabilityHub is a useful tool for this. Simply upload a file with your design and it’ll be shown to members for 5 seconds. They are then asked to write down what they remember about it.
If your website is selling shoes but the visitors remember only that you’ve been operating since 1967 or that the image carousel was annoying, it means that something is wrong.
Use UsabilityHub for evaluating pages that visitors use to enter your website: homepage, campaign-specific landing pages, product pages, category pages, etc.
First click test
First click test is another test from UsabilityHub (see how it works here). Visitors are shown a design of a website (but it can be also an e-mail or mobile app), and are asked a question about the site’s usability.
For example: where would you click to get to “sell a website?”
The tool then gives your business a snapshot of where people clicked. Visitors are only allowed to click once. This test checks if there is any friction in your navigation. It’s an eye-opener when you see visitors struggling with a task you believe is a piece of cake (because you know the website by heart already).
Observing how visitors actually use your website will give you tons of ideas on what to improve.
Write scenarios for tasks (2-3 most important user flows on your website - like placing an order, narrowing down the search on a category page etc.), and then have someone do these tasks on your live website.
My advice is to keep it small in the beginning. Test 4-5 users. If you have budget, use some tools like uxeria.com or usertesting.com. If you want to make it cheap and fast, just ask some people in the cafeteria for 30 minutes of their time and buy them coffee in return.
Remember, ask them:
- What are you looking for?
- What problems do you encounter?
- What is unclear?
- What are you thinking now?
This is the way to get as much insight as possible - not only about users behavior but also about their thought process.
To learn how to do usability testing on a budget, go through some of the articles listed by uxmyths.com.
This is simple and quick. You explore your own website and jot down possible issues. It might be related to usability, copy, images and so on.
It’s important to use a framework that will guide you towards the areas you should focus on, especially if you don’t have a lot of experience with conversion optimization. There are many guides out there:
A part of ResearchXL frameworkhttps://conversionxl.com/how-to-come-up-with-more-winning-tests-using-data/ by ConversionXL is heuristics, which cover 5 areas: relevancy, clarity, value, friction and distraction.
The LIFT Model by Widerfunnelhttps://www.widerfunnel.com/lift-model/ approaches the heuristic research in a similar way, but encourages you to take into account a factor of building urgency as well. You can also use 10 Heuristics for User Interface Design created by Jakob Nielsen back in 1995. They are more general and focused on usability only, but despite being over 20 years old they are still applicable as the ground rules of building easy-to-use websites.
User session replays
Contrary to popular heatmaps, you don’t need thousands of visitors to get insights. Watching even 20 recordings might be enough to spot the biggest problems.
It’s extremely useful to see how visitors struggle with your forms, navigation, at what point they get bored by the content and just start to scroll down, or when they are leaving your pages.
Look for technical issues and website speed
Seamless usability and well-crafted copy isn’t going to work miracles so long as your website is slow or buggy.
Quick load time is a solid foundation you should lay before going into other areas of conversion optimization.
A research published by Akamai states that 47% of web visitors expect the page to load in under 2 seconds. The takeaway here is that the faster the website, the happier (and more eager to convert) your visitors are.
The sample size in site speed reports you get in Google Analytics might be too low if your traffic is small.
Use tools like Pingdom or WebPageTest to check your site speed instead. You will be given a handful of metrics like page size, load time or performance grade. Most importantly though, the tools will provide you with a list with suggestions on what to improve to speed up the website.
Page Load Time optimization is a big topic itself and there are many detailed articles about it out there. For example you can get more knowledge about from the articles published on Kissmetrics, TruConversion or VentureHarbour.
Another low-hanging fruit might be technical problems with your website. Make sure that no matter what device or browser version your visitors are using, they can explore it and convert without encountering irritating errors or bugs.
Time to roll up your sleeves
You can stop hiding behind “I don’t have enough traffic” excuse now. You don’t need A/B testing and millions of sessions in analytics software to start optimizing your website.
The sooner you start, the more additional revenue will you make. So start now.
- Myth #22: Usability Testing Is Expensive
- 10 Usability Heuristics For User Interface Design
- Speed Is A Killer – Why Decreasing Page Load Time Can Drastically Increase Conversions
- 9 Tips To Reduce Page Load Time And Speed Up Your Website
- A/B Testing for Low Traffic Websites
- Sample size calculator - http://www.evanmiller.org/ab-testing/sample-size.html
- Five second test - http://fivesecondtest.com/
- First click test - https://usabilityhub.com/click-test
- User session recording #1 - https://www.hotjar.com/
- User session recording #2 - https://www.userreplay.com/
- User session recording #3 - https://sessioncam.com/session-replay/
- User session recording #4 - http://www.inspectlet.com/
- Page speed #1 - https://tools.pingdom.com/
- Page speed #2 - https://www.webpagetest.org/
- Browser testing #1 - https://crossbrowsertesting.com/
- Browser testing #2 - https://www.browserstack.com/
About the author:
Damian Rams helps companies increase online sales using conversion optimization. He delivers data-driven insights about website performance, which after implementation result in more leads and higher revenue.