30+ cases: how website speed affects conversion

    Website Acceleration
    Correct measurement of site conversion is not an easy task. And the connection between site speed and conversion is almost impossible at all. At least in the last 8 years, only a few companies have been able to correctly relate these two values: a little more than 30 published cases have accumulated. We carefully collected them all, laid them out by metrics, found the primary sources and offer you.

    Attention!There are three main metrics in site speed - server response time (TTFB), rendering time (DOMready), and full load time (onload). Some studies focus only on one of the metrics, and some without any of these metrics at all. Applying the research results (and proposed methods) to your site, you need to clearly understand what specific metric (and its improvement) is involved. Otherwise, the problems of your site cannot be correlated with the results.

    Site Acceleration: Google

    1. Google


    What they did: added a delay when displaying search results

    What was measured: server response time (TTFB)

    Result: 0.4 seconds of delay reduced the number of search queries by 0.76%

    Source (presentation)




    Site Acceleration: Google

    2. Google


    What they did: increased the number of results on the search page.

    What was measured: server response time (TTFB)

    Result: by 25% of search queries (ad impressions) with a delay increase of 0.5 seconds

    Source (presentation)




    Site Acceleration: Google

    3. Google


    What they did: added a delay when displaying search results.

    What was measured: server response time (TTFB)

    Result: 2% of the delay reduced the number of search queries by 2%

    Source (presentation)




    Website Acceleration: Bing

    4. Bing


    What they did: added a server delay before issuing a page with search results

    What was measured: server response time (TTFB)

    Result: 0.5 seconds of a server delay reduce profit by 1.1-1.4%

    Source (presentation)




    Website Acceleration: Strangeloop

    5. Strangeloop



    What they did: measured the relationship between the server response time and the site conversion.

    What was measured: server response time (TTFB)

    Result: 0.5 seconds delay on the server takes 1.6-1.9% of the conversion, increases the failure rate by 4.6-4 , 7%

    Source (presentation)




    Site Acceleration: Amazon

    6. Amazon


    What they did: measured the relationship between server response time and revenue

    What was measured: server response time (TTFB)

    Result: 0.1 seconds increases revenue by 1%

    Source (presentation)




    7. DoubleClick


    What they did: deleted 1 redirect in an ad transition.

    What was measured: server response time (TTFB)

    Result: a decrease in server delays by 1.5 seconds increases ad conversions by 12%

    Source (article archive)




    Website Acceleration: Facebook

    8. Facebook


    What they did: added a server delay before issuing pages.

    What was measured: server response time (TTFB)

    Result: 0.5 second delay reduces page views by 3%, 0.6 second decrease in server response time gives + 8% clicks on the tape

    Source (presentation )




    Site Acceleration: Google

    9. Google


    What they did: added a delay at different stages of page loading.

    What was measured: server response time (TTFB), rendering time (DOMready)

    Result: 0.2 seconds of delay reduced the number of search queries by 0.3%




    Website Acceleration: Bing

    10. Bing


    What they did: they made the page 5 times “heavier”

    What was measured: rendering time (DOMready)

    Result: “weighting” the page 5 times reduced clicks by 0.55%

    Source (presentation)




    Website Acceleration: Bing

    11. Bing


    What they did: accelerated the appearance of the first result on the browser screen

    What was measured: rendering time (DOMready)

    Result: acceleration by 4-18% yielded 0.7% more clicks

    Source (presentation)




    Website Acceleration: Walmart

    12. Walmart.com


    What they did: measured the speed and conversion of the site, comprehensively accelerated the site.

    What was measured: the time from receiving a response by the user to full load (onload-TTFB).

    Result: 1 second of load time gives + 2% conversion

    Source (presentation)




    Website Acceleration: Mozilla

    13. Firefox


    What they did: they did A / B testing and integrated website acceleration.

    What was measured: time to fully load the page (onload)

    Result: 2.16 seconds of load time give 15.4% of conviction (number of downloads)

    Source (article)




    Website Acceleration: Manicurshop

    14. Manicurshop


    What they did: they carried out a comprehensive acceleration of the site (vendor - WEBO Group)

    What was measured: total load time (onload)

    Result: a decrease in the load time of the site by 8.5 seconds yielded + 40% conversion

    Source (article)




    Website Acceleration: Autoanything

    15. AutoAnything


    What they did: they did a comprehensive website acceleration (vendor - Radware)

    What was measured: full load time (onload)

    Result: acceleration from 10 seconds to 5 seconds (doubled) increased conversion by 9%, average check by 11% and revenue by 13%

    Source (google cache)




    Website Acceleration: AOL

    16. AOL


    What they did: measured the relationship between page load time and page views.

    What was measured: total load time (onload)

    Result: the difference between the depth of the site’s viewing for “slow” and “fast” users reaches 60%

    Source (presentation)




    Website Acceleration: Yahoo

    17. Yahoo!


    What they did: measured the relationship between the page load time and the bounce rate.

    What was measured: total load time (onload).

    Result: 0.4 seconds of the site load time increase the bounce rate by 5-9%.

    Source (presentation)




    Website Acceleration: Soasta

    18. Soasta


    What they did: measured the relationship between page load time and conversion for mobile users.

    What was measured: total load time (onload).

    Result: with a difference in page load time of 1 second, the difference in conversion of mobile users reaches 27%, in the bounce rate - 56%

    Source (article)




    Website Acceleration: Mozilla

    19. ICTTrainingen


    What they did: they conducted A / B testing and complex website acceleration (vendor - WEBO Group)

    What was measured: total load time (onload)

    Result: reducing the website loading time by 2.5 seconds increased the conversion by 2%, and the profit per visitor by 10%.

    Source (article)




    Website Acceleration: Etsy

    20. Etsy


    What they did: they added 160 Kb of images and measured the conversion.

    What was measured: full load time (onload).

    Result: additional 160 Kb of images reduced the conversion of mobile users by 12%.

    Source (article)




    Website Acceleration: Shopzilla

    21. Shopzilla


    What they did: they conducted a comprehensive acceleration of the site.

    What was measured: full load time (onload).

    Result: a decrease in the load time of the site by 4.8 seconds increased conversion by 7-12% and page views by 25%.

    Source (presentation)

    .

    .




    Website Acceleration: Compuware

    22. Compuware


    What they did: measured the relationship between site load time and bounce rate (iPad)

    What was measured: total load time (onload)

    Result: 1 second delay in loading the site adds 2% to the bounce rate

    Source (White Paper)

    .

    .




    Site Acceleration: Glasses Direct

    23. Glasses Direct


    What they did: conducted a comprehensive website acceleration (vendor - Aberdeen Group)

    What was measured: full load time (onload)

    Result: website acceleration by 1 second gave 6.7% of conversion

    Source (article print)




    Site Acceleration: Portent

    24. Portent


    What they did: measured the relationship between the site load time and the page value.

    What was measured: total load time (onload)

    Result: website acceleration by 3 seconds increases the page value by 18%

    Source (article)

    .

    .

    .

    .




    Website Acceleration: Hotmail

    25. Hotmail


    What they did: optimized the load of advertising.

    What was measured: full load time (onload).

    Result: website acceleration by 6 seconds yielded $ 6 million per year in advertising profit, 40 million more advertising impressions per month

    Source (presentation)

    .

    .

    .

    .




    Site Acceleration: Kyle Rush

    26. Kyle Rush


    What they did: they did A / B testing and integrated website acceleration.

    What was measured: full load time (onload)

    Result: website acceleration increased conversion by 14% by 2%

    Source (article)

    .

    .




    Website Acceleration: Compuware

    27. Compuware


    What they did: measured the relationship between the site load time and the bounce rate.

    What was measured: total load time (onload).

    Result: increasing the site load time by 8 seconds increases the bounce rate by 38%.

    Source (White Paper)




    Site Acceleration: Edmunds

    28. Edmunds


    What they did: conducted a comprehensive website acceleration.

    What was measured: full load time (onload).

    Result: website acceleration by 6.5 seconds reduced failures by 4%, increased ad impressions by 3%, viewing depth by 20%

    Source (presentation)

    .

    .




    Site Acceleration: Staples

    29. Staples


    What they did: they conducted a comprehensive website acceleration.

    What was measured: full load time (onload).

    Result: website acceleration increased conversion by 10%.

    Source: ( 10% )

    .

    .




    Site Acceleration: Intuit

    30. Intuit


    What they did: they did A / B testing and integrated website acceleration.

    What was measured: full load time (onload).

    Result: website acceleration increases conversion by 3% by 1 second.

    Source (presentation)




    Site Acceleration: Google

    31. Google


    What they did: degraded image quality for users with slow Internet

    What was measured: full load time (onload)

    Result: a 30% decrease in page size gave 30% more requests from users

    Source (article)

    Also popular now: