Engagement falls off when there is a delay in experience past a certain point. Usually considered around 100ms to 150ms with extreme drop off at a second or higher. This has to go with human perception and can be measured through a/b analysis and similar.
Engagement does not get better if you go faster past that point. Past that point, you should have a richer experience, more things on the page, whatever you want. or reduce cost by spending less on engineering, certainly don't spend more money on a 'feature' (speed) that doesn't return money.
Ad networks are run on deadline scheduling. Find the best ad in 50 Ms, don't find any old ad as quick as possible.
Haven't others been involved with engagement analysis found the same?
There is one.
Engagement falls off when there is a delay in experience past a certain point. Usually considered around 100ms to 150ms with extreme drop off at a second or higher. This has to go with human perception and can be measured through a/b analysis and similar.
Engagement does not get better if you go faster past that point. Past that point, you should have a richer experience, more things on the page, whatever you want. or reduce cost by spending less on engineering, certainly don't spend more money on a 'feature' (speed) that doesn't return money.
Ad networks are run on deadline scheduling. Find the best ad in 50 Ms, don't find any old ad as quick as possible.
Haven't others been involved with engagement analysis found the same?