Conversion rate optimization

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search
A seller wants to optimize the conversion rate from "saw the advertisement" to "clicked the advertisement", as well as the conversion rate from "clicked the advertisement" to "bought the product".

In internet marketing, and web analytics conversion optimization, or conversion rate optimization (CRO) is a system for increasing the percentage of visitors to a website that convert into customers,[1] or more generally, take any desired action on a webpage.[2] It is commonly referred to as CRO.


Online conversion rate optimization (or website optimization) was born out of the need of e-commerce marketers to improve their website's performance in the aftermath of the dot-com bubble. As competition grew on the web during the early 2000s, website analysis tools and an awareness of website usability prompted internet marketers to produce measurables for their tactics and improve their website's user experience.

In 2004, new tools enabled internet marketers to experiment with website design and content variations to determine which layouts, copy text, offers, and images perform best. This form of optimization accelerated in 2007 with the introduction of the free Google Website Optimizer.[3] Today optimization and conversion are key aspects of many digital marketing campaigns. A research study conducted among internet marketers in 2014, for example, showed that 59% of respondents thought that CRO was "crucial to their overall digital marketing strategy".[4]

Conversion rate optimization shares many principles with direct response marketing – a marketing approach that emphasizes tracking, testing, and on-going improvement. Direct marketing was popularized in the early twentieth century and supported by the formation of industry groups such as the Direct Marketing Association, which formed in 1917.[5]

Like modern day conversion rate optimization, direct response marketers also practice A/B split-testing, response tracking, and audience testing to optimize mail, radio, and print campaigns.[6]

Statistical significance[edit]

Statistical significance helps understand that the result of a test is not achieved merely on the basis of chance. It helps understand the confidence and risk level of a test.[citation needed] Frequently, when marketers study a lift in an ad campaign, they discover customer behavior is not consistent. Online marketing response rates fluctuate widely from hour to hour, segment to segment, and offer to offer [7].

This phenomenon can be traced to the difficulty humans have separating chance events from real effects. Using the haystack process, at any given time, marketers are limited to examining and drawing conclusions from small data samples. However, psychologists (led by Daniel Kahneman and Amos Tversky) have documented tendencies to find spurious patterns in small samples to explain why poor decisions are made. Statistical methodologies can be leveraged to study large samples, mitigating the urge to see patterns where none exist.

These methodologies, or "conversion optimization" methods, are then taken a step further to run in a real-time environment.[citation needed] The real-time data collection and subsequent messaging increases the scale and effectiveness of the online campaign.[citation needed]

Reaching a statistically significant result in itself is not enough. Conversion optimization practitioners must ensure that their sample size accounts for important variables. For example, a test may appear statistically significant well before seasonal factors (time of day, day of week, time of year) have been adequately reflected in the data sample. One variation may appeal to one season more than others and ultimately misguide the result.[citation needed]

It is equally important to understand how various segments affect tests and results. Different user segments (e.g. device type, location, new vs. returning visitor) will respond differently to each variation.[citation needed]. Analyzing results without accounting for different segments can cause a significant improvement for one segment; or many variations can offset poor results for another segment. For example, uplift in desktop conversion-rate could offset a decreased conversion-rate on mobile devices. In this instance, only the desktop version should be declared a ‘winning’ test.[citation needed]


Conversion rate optimization seeks to increase the percentage of website visitors that take a specific action (often submitting a web form, making a purchase, signing up for a trial, etc.) by methodically testing alternate versions of a page or process.[citation needed] In doing so, businesses are able to generate more leads or sales without investing more money on website traffic, hence increasing their marketing return on investment and overall profitability.[8]

A conversion rate is defined as the percentage of visitors who complete a goal, as set by the site owner. Some test methods, such as split testing or A/B testing, enable one to monitor which headlines, copy, images, social proof elements,[citation needed] and content help convert visitors into customers.

There are several approaches to conversion optimization with two main schools of thought prevailing in the last few years.[citation needed] One school is more focused on testing to discover the best way to increase website, campaign, or landing page conversion rates. The other school is focused on the pretesting stage of the optimization process.[citation needed] In this second approach, the optimization company will invest a considerable amount of time understanding the audience and then creating a targeted message that appeals to that particular audience. Only then would it be willing to deploy testing mechanisms to increase conversion rates.

Elements of the test focused approach[edit]

Conversion optimization platforms for content, campaigns, and delivery consist of the following elements:

Data collection and processing[edit]

The platform must process hundreds of variables and automatically discover which subsets have the greatest predictive power, including any multivariate relationship. A combination of pre- and post-screening methods is employed, dropping irrelevant or redundant data as appropriate. A flexible data warehouse environment accepts customer data as well as data aggregated by third parties.

This means it's essential to ensure the data is as 'clean' as possible, before undertaking any data analysis. For example, eliminating activity from bots, staging websites, or incorrect configurations of tools such as Google Analytics.

Data can be numeric or text-based, nominal or ordinal. Bad or missing values are handled gracefully.

Data may be geographic, contextual, frequential, demographic, behavioral, customer based, etc.


After data collection, forming a hypothesis is the next step. This process forms the foundation of why changes are made. Hypotheses are made based on observation and deduction. It is important that each hypothetical situation be measurable. Without these no conclusions can be derived.

Optimization goals[edit]

The official definition of "optimization" is the discipline of applying advanced analytical methods to make better decisions. Under this framework, business goals are explicitly defined and then decisions are calibrated to optimize those goals. The methodologies have a long record of success in a wide variety of industries, such as airline scheduling, supply chain management, financial planning, military logistics and telecommunications routing. Goals should include maximization of conversions, revenues, profits, LTV or any combination thereof.

Business rules[edit]

Arbitrary business rules must be handled under one optimization framework. Using such a platform entails that one should understand these and other business rules, then adapt targeting rules accordingly.

Real-time decision making[edit]

Once mathematical models have been built, ad/content servers use an audience screen method to place visitors into segments and select the best offers, in real time. Business goals are optimized while business rules are enforced simultaneously. Mathematical models can be refreshed at any time to reflect changes in business goals or rules.

Statistical learning[edit]

Ensuring results are repeatable by employing a wide array of statistical methodologies. Variable selection, validation testing, simulation, control groups and other techniques together help to distinguish true effects from chance events. A champion/challenger framework ensures that the best mathematical models are deployed always. In addition, performance is enhanced by the ability to analyze huge datasets and to retain historical learning.

See also[edit]


  1. ^ Kahled, Saleh and Shukairy, Ayat (2011). Conversion Optimization: The Art and Science of Converting Prospects into Customers, p. 2. O'Reilly Media, Sebastopol. ISBN 978-1-449-37756-4.
  2. ^ Ash, Tim and Page, Rich and Ginty, Maura (2012). Landing Page Optimization, p. 13. Wiley & Sons, Indianapolis. ISBN 9780470610121.
  3. ^ Page, Rich (2012). Website Optimization: An Hour a Day, Wiley & Sons, Indianapolis. ISBN 978-1-118-19651-9.
  4. ^ "Conversion Rate Optimization Report 2014". London: Econsultancy. 2014. Retrieved 2 December 2014. Cite journal requires |journal= (help)
  5. ^ dma. "About the Direct Marketing Association".
  6. ^ "Tested Advertising Methods".
  7. ^ Dvir, Nim; Gagni, Ruti (2018). "When Less Is More: Empirical Study of the Relation Between Consumer Behavior and Information Provision on Commercial Landing Pages". Informing Science: The International Journal of an Emerging Transdiscipline. 21: 019–039. doi:10.28945/4015. ISSN 1547-9684.
  8. ^ "You Should Test That".