Conversion Rate Optimization Flashcards

Question 1

What is conversion rate in its simplest form?

Accepted Answer

Conversion rate is the number of visitors who complete a target action divided by the total number of visitors in the same period, usually expressed as a percentage. For example, if 150 of 5,000 visitors place an order, the conversion rate is 3 percent.

Question 2

What counts as a conversion goal in web analytics?

Accepted Answer

A conversion goal is any measurable action that delivers business value, such as a completed purchase, a form submission, a phone call, or a download. Macro conversions are major outcomes like sales, while micro conversions are smaller steps such as adding an item to a cart or subscribing to a mailing list.

Question 3

What is a conversion funnel and why do analysts use it?

Accepted Answer

A conversion funnel is a sequence of steps a visitor must take to reach a goal, such as viewing a product, adding it to a cart, entering shipping details, and paying. Analysts use it to count how many users reach each stage and where they leave. Large drops show where to look first.

Question 4

What is the difference between a click-through rate and a conversion rate?

Accepted Answer

A click-through rate measures the share of people who saw a link or ad and clicked it, while a conversion rate measures the share who later completed a goal. A campaign can earn many clicks yet few conversions if the landing page fails to deliver what the ad promised.

Question 5

Why can a conversion rate from a small number of visitors be misleading?

Accepted Answer

A small sample produces rates that swing widely from random chance alone. If two visitors convert out of ten, the rate is 20 percent, but a few more or fewer purchases would change it dramatically. Analysts should wait for enough visitors and conversions before drawing conclusions, since early numbers can mislead.

Question 6

What is a baseline conversion rate and why does it matter?

Accepted Answer

The baseline conversion rate is the current performance of a page, flow, or segment before any change is made. It gives every test a reference point, so the team can measure whether a variant actually improves results. Without a reliable baseline, teams cannot estimate how many visitors an experiment needs.

Question 7

What is average order value and how does it relate to conversion?

Accepted Answer

Average order value is the mean revenue generated per completed order, calculated by dividing total revenue by the number of orders. It is separate from conversion rate, so a change can raise one while lowering the other. Track both, since a test that boosts conversions but shrinks order size may cut total revenue.

Question 8

Why is revenue per visitor a useful companion metric?

Accepted Answer

Revenue per visitor divides total revenue by total visitors, combining conversion rate and order value into one number. It shows the overall value of traffic, so a page with fewer buyers can still win if those buyers spend more. Comparing variants on it helps avoid misleading wins.

Question 9

What is the purpose of an analytics tag or tracking event?

Accepted Answer

A tracking event is a recorded signal that a specific user action occurred, such as a button click, form submission, or page view. Tags are the snippets of code that send these signals to an analytics platform. Clear event naming and consistent parameters let analysts build funnels and segments reliably.

Question 10

Why do analysts segment conversion data by device, source, or audience?

Accepted Answer

Segmenting splits overall results into groups that behave differently, such as mobile and desktop users or paid and organic visitors. An average can hide a problem in one group, so a page that looks flat overall may convert well on desktop and poorly on phones. Segments also suggest which audiences need tailored tests.

Question 11

What is the difference between a session and a unique user?

Accepted Answer

A session is one continuous period of activity by a visitor. A unique user is an individual person or device counted once within a time window, regardless of how many sessions they start. One user can produce several sessions, so conversion rates depend on the unit reported.

Question 12

How does a cookie or identifier affect conversion tracking accuracy?

Accepted Answer

Analytics tools usually identify returning visitors with cookies or similar identifiers stored in the browser. When users clear storage, block tracking, or switch devices, one person may appear as several visitors, which can distort counts. Consent rules and browser privacy settings can reduce the data available, so reported rates are estimates.

Question 13

What is data sampling and why can it matter?

Accepted Answer

Data sampling means an analytics tool analyzes a subset of records instead of the full dataset to produce reports faster. Estimates from a sample can differ from full counts, especially for small segments or narrow date ranges. When precision matters, analysts should check whether sampling is applied and use unsampled data.

Question 14

Why should analysts check tracking quality before running a test?

Accepted Answer

Tracking quality checks confirm that goals fire once, on the correct pages, and for both control and variant. Broken or duplicate events can make a winning variant look better or worse than it really is. Auditing before launch protects the decision, since lost data rarely can be recovered.

Question 15

What does it mean if a test's traffic split is uneven?

Accepted Answer

An uneven split means the variants receive visitors in proportions different from the planned ratio, such as 60 percent and 40 percent instead of 50 and 50. This is called sample ratio mismatch and usually signals a technical fault. Results from such a test are unreliable until the cause is found and fixed.

Question 16

Why is it important to define a test's primary metric before launching?

Accepted Answer

The primary metric is the single outcome that will decide whether the variant wins. Naming it in advance prevents teams from searching through many measures after the data arrives and choosing whichever one looks favorable. Secondary metrics can still be reviewed, but they should inform follow-up questions rather than override the planned decision.

Question 17

What is an A/B test and how does it work?

Accepted Answer

An A/B test randomly assigns visitors to either a control version or a variant, then compares a goal metric between the two groups. Random assignment spreads differences in visitor type evenly across groups, so any gap in results is more likely caused by the change itself.

Question 18

Why does every conversion experiment need a control group?

Accepted Answer

A control group is the portion of visitors who see the current experience, not the new idea being tested. It provides the baseline against which the variant is compared, so teams can separate the effect of the change from background trends. Without a control, any improvement could reflect changes in who was visiting.

Question 19

What does statistical significance mean in a test result?

Accepted Answer

Statistical significance indicates that an observed difference would be unlikely to appear if the variant truly had no effect. It is usually judged against a threshold chosen before the test begins. Significance does not prove the size of the effect is important, and it does not guarantee the result will repeat.

Question 20

What does a p-value actually tell an analyst about a result?

Accepted Answer

A p-value is the probability of seeing a difference at least as large as the observed one if there were truly no difference between variants. A small p-value suggests the data are hard to explain by chance alone. It does not give the probability that the variant is better.

Question 21

What is statistical power and why does it matter?

Accepted Answer

Statistical power is the probability that a test will detect a real effect of a given size when that effect exists. Low power means a test may miss genuine improvements, leading teams to abandon good ideas. Power depends on sample size, expected effect size, and the chosen significance level.

Question 22

How does sample size relate to detecting a small effect?

Accepted Answer

Detecting a small difference generally requires a larger sample than detecting a large one, because small effects are easier to hide in normal random variation. Teams often underestimate how many visitors they need, which leads to tests that end early with inconclusive results. Calculating the required sample before starting helps set realistic timelines.

Question 23

What is a Type I error in testing?

Accepted Answer

A Type I error is a false positive, meaning the test declares a winner when the variant actually has no real effect. Its probability is controlled by the significance threshold, so a stricter threshold lowers the chance of this mistake. Stopping a test as soon as results look good increases false positives.

Question 24

Why do analysts often commit to a fixed test duration?

Accepted Answer

A fixed duration covers complete weekly cycles, so weekday and weekend behavior are both represented in the data. Stopping as soon as results look significant allows random fluctuations to make a weak variant appear strong. Committing to a sample size or runtime in advance protects the decision from this bias.

Question 25

What is friction and how does it affect conversion?

Accepted Answer

Friction is any element that slows or complicates a user's progress toward a goal, such as unclear labels, forced account creation, long forms, or unexpected costs. Each point of friction raises the chance that a motivated visitor will give up. Reducing friction often improves conversion.

Question 26

How does visual hierarchy influence which action visitors take?

Accepted Answer

Visual hierarchy is the arrangement of size, color, contrast, and spacing that tells the eye where to look first. A clearly dominant primary button draws attention to the main action, while competing links, bright banners, or equal-weight options can split attention. Designers use hierarchy to guide visitors toward the goal.

Question 27

Why should clarity of value matter on a landing page?

Accepted Answer

Visitors decide within seconds whether a page is relevant, so the headline and supporting copy should state what the offer is and who it serves. Vague promises slow that recognition. Specific benefits, matched to the traffic source, help visitors quickly see they are in the right place.

Question 28

What is social proof and when can it help conversion?

Accepted Answer

Social proof is evidence that other people have chosen or endorsed an option, such as customer reviews, ratings, or counts of satisfied users. It can reduce uncertainty when a visitor is unfamiliar with a brand or a product's quality is hard to judge from the page. Its effect depends on credibility, so fabricated testimonials can damage trust.

Question 29

How does reducing perceived risk affect a purchase decision?

Accepted Answer

Perceived risk covers worries about money, privacy, returns, and whether a product will work as expected. Trust signals such as clear return policies, secure payment badges, and transparent pricing address these concerns directly. Because hesitant visitors often leave without saying why, removing doubt at the decision point can lift completion.

Question 30

Why should mobile layout receive special attention in conversion work?

Accepted Answer

Many visitors arrive on phones, where small screens magnify problems like tiny tap targets, hidden buttons, and slow-loading images. Forms are especially difficult on mobile, so fields that auto-fill or use the right keyboard type can save real effort. Testing on real phones catches problems that desktop previews often miss.

Question 31

What role does page speed play in conversion?

Accepted Answer

Slow pages lose visitors before they can read the offer, and each delay increases the chance of abandonment. Speed affects both perceived quality and practical usability, especially for users on slower connections or older devices. Teams should measure load and responsiveness, then prioritize fixes that remove the largest delays for real visitors.

Question 32

What is cognitive load and why does it matter for forms?

Accepted Answer

Cognitive load is the mental effort required to understand and complete a task. Forms with many fields, unclear instructions, or confusing error messages increase that effort, so people abandon them partway through. Asking only for necessary information, grouping related fields, and showing errors clearly all reduce load and tend to improve completion rates.

Question 33

What steps should a team follow to run a conversion test?

Accepted Answer

A sound process begins with research into user behavior, then forms a clear hypothesis, defines the primary metric, and calculates the required sample. The team builds the variant, checks tracking, runs the test for a predetermined period, and analyzes results against the plan. Documenting each step helps teams learn from wins and losses.

Question 34

What makes a good hypothesis for a conversion test?

Accepted Answer

A good hypothesis states a proposed change, the reason it should work based on evidence, and the outcome expected. For example, shortening the checkout form may increase completed orders because visitors will face fewer fields. This structure keeps the test focused and makes the learning useful, whether the variant wins or loses.

Question 35

Why should a team avoid changing many elements in one variant?

Accepted Answer

When several elements change at once, the team cannot tell which change caused a result. A bundled variant may win overall while one change hurts and another helps, hiding useful lessons. Testing a focused change, or using a structured multivariate design with enough traffic, keeps conclusions clear and actionable.

Question 36

How should teams interpret a result that is significant but tiny?

Accepted Answer

A statistically significant lift can still be too small to justify the cost of building, maintaining, and explaining a change. Teams should compare the confidence interval with the minimum effect worth acting on. If the likely range of gains is trivial, the change may not deserve a place in production.

Question 37

What is a confidence interval and how should it be read?

Accepted Answer

A confidence interval gives a range of plausible values for the true difference between variants, based on the observed data and a chosen confidence level. A wide interval signals uncertainty, while a narrow one suggests more precise estimates. Read it as a range of possibilities rather than a single exact number.

Question 38

Why do analysts check for novelty effects after a test?

Accepted Answer

A novelty effect happens when returning visitors react strongly to a change simply because it looks different, and the boost fades over time. Early results can therefore overstate long-term impact. Analysts often review performance across weeks, and compare new and returning visitors, to judge whether gains persist.

Question 39

How should a team document and share test results?

Accepted Answer

A useful record states the hypothesis, the audience, the dates, the primary metric, the result with its uncertainty, and the decision made. Sharing it in a central repository prevents teams from repeating failed ideas and lets others build on successful ones. Include notes on implementation problems, because they often explain surprising outcomes later.

Question 40

What should a team do after a test concludes, whatever the outcome?

Accepted Answer

The team should implement the decision according to the pre-set rules, whether the variant wins, loses, or shows no clear difference. Monitoring the live experience afterward confirms that results hold in production and that no unexpected problems appear. Every conclusion, including a null result, should feed the next round of research and hypotheses.

Conversion Rate Optimization Flashcards

1Core Concepts and Metrics

2Measurement and Tracking Foundations

3Experiment Design and Statistics

4User Experience and Persuasion

5Running and Applying Tests

Learning Resources