impact of speed of application on conversion rates

[starec]

I am exploring a possibility to significantly increase the speed of product search by using a local caching system.

However, I have little experience in what impact to expect. Does anybody here have real measurements of a change like this? I.e., if I reduce the time of a product search from 20 to 10 seconds, will the conversion improve just a bit, or will it really go up significantly?

[StuWhite]

Even 10 seconds sounds like a very long time to wait for search results. I'm suprised that anyone at all would hang around for 20 seconds.

If your database access is really that slow (and can't be improved) then I'd be looking at implementing something like Google site search instead.

[starec]

Now it takes at times even more than 20 seconds because it's an external data source. With local caching I can get it down to 5-10 seconds, depending on the processing that has to be done before showing the results.

[justgowithit]

20 seconds to five seconds....? It's a no-brainer. Of course conversion will improve.

[jsinger]

But remember that most shoppers (on all but the biggest sites) find products by drilling down, not by using a search form.

If your search is that slow, work also on alternaive navagation.

Yes, 20 seconds is ridiculous.

[bmcgee]

That's a blanket statement that doesn't apply to all types of sites/products.

[vincevincevince]

This is a blanket statement which applies to all website features:

20 seconds is too long to ask someone to wait for search results.

[limoshawn]

I have to agree, I won't wait 20 seconds for a video load and I sure won't wait that long for search results. In fact I can't think of anything on the net I would wait 20 seconds for.

[jsinger]

I can't think of anything on the net I would wait 20 seconds for.

Early on, porn took far longer to load. Didn't stop many. :)

[Laker]

Early on, porn took far longer to load. Didn't stop many. :)

That was before Viagra.

[bmcgee]

Perhaps I was not clear. When referring to the "blanket statement" it was that users drilldown and don't use search functionality on websites. Findings alternative navigation is NOT an option sometimes, unless you want to alienate the vast majority of your visitors in those industries.

Of course, search that is performing that poorly needs to be revisited and improved significantly. But to suggest to throw search out the window could be a deathwish in some cases.

[jsinger]

Read what I said:

most shoppers (on all but the biggest sites) find products by drilling down, not by using a search form.

Certainly search forms are used to some degree. They are vital on some sites. But many work terribly. Look up "baseball", "baseballs", "base ball", "base balls" on smaller sporting goods sites. You'll get different results with each and probably nothing with one or two. Shoppers have learned to navigate by better means.

We have a search form on nearly every page of our site and only a small portion of our visitors do even one form search.

The subject has come up before on WebmasterWorld

[bmcgee]

Read what I said:
I said your experience is not the way of the world in all industries. In some industries, people won't drilldown, they will search. It is just the nature of the products and of the demographics.

Just because most do it on most websites, doesn't mean there aren't some types of sites where nearly ALL search instead of drilldown.

We don't know the OP's site or industry. So blindly tell him to get a different way of navigation for his site rather than solve the issue is inappropriate.