Conversion rate optimization lives and dies by measurement. Without analytics, you’re guessing on what works. With the wrong analytics focus, you’re measuring with confidence and learning nothing useful.
The difference between CRO programs that achieve consistent improvement and those that spin their wheels may often be what gets measured and how those measurements are used to make decisions. Not all of the numbers in your analytics dashboard are equally important. Some of the metrics show a real insight of user behavior and conversion barriers. Others just look good on reports without helping to act in meaningful ways.
Understanding what metrics are actually driving growth and how to use them and how it transforms CRO from random testing to systematic improvement.
Before jumping to particular metrics, the way measurements are taken is in need of attention. Effective CRO analytics begins not by asking what you can count, but what you’re trying to learn.
Every metric should relate to a question you’re trying to answer. What’s making visitors leave this page? Why are conversion rates lower for mobile users? Which traffic sources generate customers that actually purchase? Metrics with no questions behind them are noise rather than signal.
This means to resist the temptation to track everything available. Today’s analytics tools are capable of measuring hundreds of data points. Most of those data points will not help you improve conversions. The practice of paying attention to the metrics that answer specific questions helps keep your attention where it is needed.
The difference between leading and lagging indicators is of enormous importance here. Lagging indicators such as overall conversion rate tell you the outcome of your efforts. Leading indicators similar to form field abandonment or scroll depth tell you what’s happening along the way. Both are important but leading indicators provide guidance for action while lagging indicators measure results.
Conversion rate is the obvious starting metric however, how you calculate and segment your conversion rate determines whether that number gives or takes away your insight.
Overall site conversion rate gives you a single number which you can easily track over time. But this aggregate masks significant variation. Traffic from different sources go into conversion at radically different rates. Mobile and desktop visitors behave differently. Different people who are new visitors and return visitors have different patterns.
Segmented conversion rates show where there actually are problems and opportunities. Breaking conversion down by device, traffic source, landing page and visitor type reveals to you where performance is falling short, specifically. A site-wide conversion rate of 3% might conceal the fact that desktop is converting 5% and mobile is converting 1.5%, a difference that points to exactly where the optimization effort should be directed.
Micro-conversions should be tracked along with final conversions. These in-between actions – email signups, add to cart clicks, video views, scroll milestones – indicate movement through your funnel even when visitors don’t complete the ultimate conversion. Tracking micro-conversions shows where the journey fails.
Beyond counting conversion, behavioral analytics tells you how visitors interact with your pages. These metrics can help you understand why conversions are happening or not.
Bounce rate lets you know the percentage of visitors who leave after seeing only a page on your website. High bounce rates for landing pages indicate some sort of mismatch between what the visitors were led to expect and what they got. But context is important – a high bounce rate on a blog post could be acceptable if the reader got what it needed, but the same for a product page should be seen as problematic.
Time on page and depth of scrolling indicate level of engagement. Are visitors really eating your content or are they leaving immediately? Do they scroll past your headlines to get to important information underneath? These metrics help you to understand if the page content is connecting or it is falling flat.
Exit rate by page – finds out where visitors most often leave your site. Pages with abnormally high exit rates may have friction or confusion stopping them from taking the journey. Finding these pages point optimization efforts towards high impact opportunities.
Click patterns and heatmaps indicate exactly what visitors interact with on every page. Are they clicking on your call-to-action buttons or what? Do they notice important information to them or scroll past it? Visual behavioral data shows attention patterns that cannot be obtained from raw numbers.
Looking at your conversion path as a funnel with different stages from entering to converting helps you to systematically analyze the different points where visitors drop off.
Step by step progression rates indicate what percent of visitors move from one stage to the next. If you have a five-step checkout, knowing that 80% of visitors will get past step one but only 40% will go on to step two tells you where exactly the leakage point is occurring.
Form analytics details observe specific conversion points. Which form fields make people abandon? How much time do the visitors spend on each field? Do specific combinations of fields relate to completion/dropout? This granular view uncovers specific points of friction that aggregate form conversion rates miss.
Path analysis provides an understanding of the paths visitors actually take, and not the paths you designed. Do visitors follow the flow or do they go off course? Which alternative pathways are conversion pathways and which are dead ends? Understanding actual behavior from assumed behavior often reveals surprising possibilities for optimization.
A/B testing lies at the very heart of CRO and proper measurement is what defines whether test results actually mean anything.
Statistical significance tells you if differences between variants that are observed are real differences in performance, or if they are because of random variation. Running tests until you have statistically significant results helps you not to draw conclusions based on noise. Most practitioners try to get at least 95% confidence before calling winners.
Sample size requirements should be calculated before tests are started. How many visitors does each variant require to be able to draw reliable conclusions? This is something that, if we can understand up front, helps us to not stop tests too early, or run them on too long.
Segmented test results – Sometimes it is found that a variant works differently for different groups of users. A change could be beneficial to mobile visitors and detrimental to desktop visitors. Checking results by segment helps not to miss these nuances for aggregate data.
Unintended consequences are caught during tests by secondary metrics. A change that may improve one conversion measure may negatively affect another one. Tracking multiple relevant metrics throughout tests ensures you’re not trading one problem for another.
Not all traffic is created equal and knowing which traffic sources are delivering the visitors who actually convert helps to guide acquisition strategy and onsite optimization.
Conversion rate by source indicates which channels are responsible for generating visitors likely to convert and those that are generating mostly casual browsers. This helps us know both where to invest the acquisition budget, and how to optimize experiences for those visitors from different sources.
Multi-touch attribution recognises that conversions are likely to have a multitude of touchpoints, not just one visit. Understanding which combinations of channels lead to conversion ensures that last-touch sources are not over-credited and earlier influences under-valued.
Returning visitor analysis is a comparison of behavior and conversion patterns between new and returning visitors. The differences provide opportunities of how to optimize initial experiences and bring visitors back for later engagement.
Having the right metrics is not a lot of use without processes in place to transform measurement into action.
Regular review cadence ensures analytics inform decisions actually. Weekly reviews of important metrics catch problems that might be emerging quickly. Monthly deeper diving to identify trends and opportunities. Quarterly analysis evaluates whether your overall CRO approach is working.
Documentation of insights helps to preserve learning over time. What did each test teach you? What hypotheses were supported or refuted by the data? Building institutional knowledge helps avoid doing things that have failed, as well as helping the new team members understand what’s already been done.
Hypothesis-driven testing makes the link between measurement and action direct. Each test should have a clear hypothesis – that is, a specific prediction about what will happen and why. Measurement then affirms or negates that hypothesis, thereby creating a genuine understanding as opposed to just finding random winners.
Analytics brings CRO from guesswork to the process of systematic improvement, but only if it is focused on metrics that actually drive insight and action. Segmented conversion rates show where there are problems. Behavioral metrics, why visitors do or don’t convert. Funnel analysis helps you to identify specific friction points. Testing metrics are a way to ensure that the differences we see really exist. Source attribution attributes the quality of traffic to conversion performance. Building the capability for regular review processes and hypothesis-driven testing ensures that measurement translates into consistent improvement.
Segmented conversion rates, funnel progression rates, behavioural metrics such as bounce rate and scroll depth and form analytics are usually the most actionable insights. The particular metrics that it is important to measure depend on your business model, and what your conversion bottlenecks are.
Statistical significance – usually significance is set at 95% confidence or greater means that the differences that are found probably reflect real differences in performance and are not being caused by random variation. Calculate the required sample sizes before you start testing and don’t stop tests until you reach these thresholds.
Lagging Indicators such as overall conversion rate measure results after they have occurred. Leading indicators such as form field completion rates or scroll depth indicate behaviour as it happens. Leading indicators – Help you to predict and influence outcomes; Lagging indicators – Confirm whether your efforts worked.
Aggregate conversion rates conceal significant variation among different user groups, devices, traffic sources and page types. Segmentation shows specifically where there are conversion problems allowing for targeted optimization rather than unfocused changes.
Weekly reviews of key metrics catch new emerging issues quickly. More in-depth analysis is done on a monthly basis and trends and patterns are identified. Quarterly evaluation to determine if your overall CRO strategy is having results. The right cadence depends on your traffic volume and testing velocity.
C – 81C, Sector – 8, Noida 201301, India
1025 Beamon Drive, Franklin 37064 Tennessee, USA
Sveavagen 34 111 34 Stockholm, Sweden
Copyright ©ACE. All rights reserved.
