17 replies to this topic

[EGOL]

When you go to a large website as a return visitor do you search or do you navigate to your destination?

I am trying to understand my visitors..... 25% of the visitors who land on the homepage go straight to the search and 1/2 of those searching arrived by a direct route (bookmark or typed in the domain).

What could that indicate? Is that good or should it cause concern that the linkage pathways are not meeting the needs?

Opinions, personal habits, speculation and authoritative replies all appreciated. Thanks!

Edited by EGOL, 17 December 2007 - 09:00 PM.

[joedolson]

Well, that depends on the search and/or the navigation.

It's not uncommon that I'll have an existing bookmark which may not go to the page I actually need - I'll use that bookmark as a shortcut to get to the site, and then find my actual goal.

Whether I use the search or the navigation depends on each site individually. On this site, for example, I'll rarely use the search --- because it's just not a very good search facility. If a site has great search; I search. If it has great navigation, I navigate.

It all depends on circumstances, really.

[kensplace]

Same here, depends on the site..

For example with ebay, I usually use search to find items, but navigate to various areas such as myebay etc.

On forums, I tend to just use 'view new posts' and only use search sparsely (as usually the search on forums is not that good).

Overall, if the search works well, I would tend to use it over navigation - if it saved time...

[EGOL]

Thanks for your replies! I have been looking at the queries submitted through the search. Most of them are very basic - one word - so it looks like they are trying to see the generic content of the site rather than going for a specific page.

For a site like wikipedia you almost have to use search to find anything. On ebay you can navigate if you want general categories or search to find specific items.

I need to think about this.

I have a search box at the top of the homepage and one at the bottom... I was really surprised to see how many searches were being performed in the bottom box - about 5% of visitors (25% were using the search at top of page). Before I put the search box on the bottom of the page, those folks might have been leaving the site.

[swainzy]

I usually go straight for the search box and use quotes if I use more than one word.

Edited by swainzy, 18 December 2007 - 12:39 AM.

[ccera]

Interesting that you mention getting mostly one word queries. I tend to get frustrated with site searches in general because they have no real "expected" behavior - there are just too many variances from site to site. With a search engine, one knows to use quotes, boolean expressions, etc. With a site search, it is hit and miss whether those things will work. Often there is no advanced search option, or it is only available once one has reached the first set of search results. Due to that frustration, I often use one word, as specific as I can make it, and if it returns less than 50 or even 100 search results, I skim for what I need. If others do the same sort of thing, it might explain the "basic" searches you described.

Maybe in general we try to keep the search box obvious yet unobtrusive (don't want it to detract from the aesthetics of the page, etc.), but now that I think about it I am always grateful to find one with a short description of the search logic used. It seems this all goes back to not making the user think.

There are sites (and I have a particular one in mind which I won't shame here) which have pages that I have only ever managed to reach through Google. Neither the navigation nor the site search will get me there; as far as I can determine the pages are completely orphaned. If not for these pages containing completely unique content (which just happens to be coveted by my little one in the case of my nameless example), I wouldn't bother. If that's not describing your site, you're probably doing okay.

CiCi

[AbleReach]

If I know the site search has worked in the past, and the site has a range of references for my general search terms, I am likely to be happy with following any link to the site and then using search.

A good site search can be faster than following navigation links, and give me access to annotations/blurbs. No matter how good the site nav is, it can't do annotated links like site search can!

[Respree]

....25% of the visitors who land on the homepage go straight to the search ....Opinions, personal habits, speculation and authoritative replies all appreciated.

This one is speculative.

Hard to say. One might assume that 75% are navigating, but you don't say what percentage come to the home page and immediately leave (use neither search nor navigation). Let's say for the moment, your homepage bounce rate is in the single digits. That would mean that nearly 3 out of every 4 people choose to navigate.

Is that good or should it cause concern that the linkage pathways are not meeting the needs?

I had the opposite reaction. To me 3 out of every 4 people navigating would lead me to the conclusion your navigational structure is good. I think it were confusing or not obvious, most would choose to search rather than navigate.

I think the size of the site or perception of the size of the site has a lot to do with whether people choose to navigate or search. With 150,000 product pages, my site tends to be a monster to navigate trying to fit them into about 1,500 categories (which includes sub categories and sub-sub-categories) and would guess my stats are the opposite of yours (that more people search, rather than navigate).

With a smaller site with less pathways, it might be easier to navigate.

To me, the more important question than the split between navigation and search is "did they find what they were looking for?"

Again, the above is speculation.

[bwelford]

I use a Google custom search on my websites and it's excellent. At least doing that they can't try to impose their distorted views of "Do No Evil" on the keyword searches.

[EGOL]

Respree: To me, the more important question than the split between navigation and search is "did they find what they were looking for?"

There is where the answer can be found.... Now how can I get to it?

... Maybe I should do analytics on the two populations... those who search and those who click something and compare the pageviews and time on site to see if there are differences.

ccera: I tend to get frustrated with site searches in general because they have no real "expected" behavior - there are just too many variances from site to site....... I often use one word, as specific as I can make it, and if it returns less than 50 or even 100 search results, I skim for what I need. If others do the same sort of thing, it might explain the "basic" searches you described.

I have a google search box so they get high quality results. I think that your method of searching matches what is being done on my site... most of the one word queries will only yield a dozen results on my site ... or they will yield a thousand and the pages with the best navigation show first in the results. So for this site, the one word queries find what they want - if my site has it. If not they can switch the search box to search the web instead of my site.

===========

These comments have been extremely helpful. Thanks!! ... still open for more comments from anyone who has them.

[bragadocchio]

Lots of possibilities are implied in the use of a site search when arriving at a page.

Here are a few:

1. The visitor is a return visitor, and the navigational search is a way of quickly finding a previously visited page, and related pages quickly.

2. The visitor is a first time visitor, and wants to quickly explore the scope of coverage of the site on a specific topic.

3. The visitor, new or returning, is accustomed to a search paradigm, and enjoys searching through a site that has an effective site search

4. The visitor may have looked at the navigation first, and may not have seen what they were looking for, or decided that they might find more using the site search.

5. The visitor isn't quite sure what they are looking for, and uses site search as a method of discovering what is available on the site (slightly different from number 2).

Putting a site search at the bottom of the page sounds like it was a pretty good idea.

[Respree]

There is where the answer can be found.... Now how can I get to it?

An excellent question. I don't know the answer, but if you come up with a way, post it here.

One thing I found helpful was to have a little script which throws all of your search terms into a simple flat file. You can then import that into a database program (like Access) to find out the number of occurences. Maybe take a look a some of the high volume search phrases and then enter then into your search engine. You know your content better than anyone else. If what you expect to see is displayed in your SERPs, I think that's about the best one could hope for.

Let's try turning the question around. If 'you' were on your site and pretending your knew nothing about it, which would be 'your' preference -- search or navigation? Is your personal choice consistent with the majority of your visitors who choose to navigate?

Edited by Respree, 18 December 2007 - 02:42 AM.

[eKstreme]

There is where the answer can be found.... Now how can I get to it?

A simple rough metric is: what's the bounce rate of the search results? To rephrase, how many searchers click on a result? You can go deeper and ask, how many searchers *return* to the site (which means you answered their question).

Some experiments for you to try:

1. Put a "recent searches" list on your home page and monitor its usage.

2. Put a tag cloud of the most popular topics on your site. Hack the CMS to generate that.

3. Put a search box prominently on the home page (if it's not already there).

Pierre

[Autocrat]

Do you know the individual search terms being used?
If so, you can easily examine them and see if there is a trend to them.
You can also look and see if the contents are easily reachable through the Navigation (within one or two clicks), and if the required content is easily noticeable (main page content, or only reference!).

Once you see what peopole are looking for and what they aregetting, you should havea better understanding as to why they are using the navigation.

If at all possible, and not in conflict with your setup, setting lengthy sessions may be of use to you as well (or cookies?)... so that you can track certain users over time and watch their behaviour. That will help[ indicate whether they migrate from one form of location of content to another, or if they get more specialised/specific with a particualr way of getting to content.

A third option is to use some conditional code to display a little reqest for feedback to users that fit yuour profile... and see if they are willing to tell you what they are doing and why.

[Wit]

Umm, Egol, do these search phrases have lots of HTML tags in them, perchance? Just a long shot.... :angel:

[EGOL]

I am getting LOTS of takeaways from your comments... This is why I post here...

bragadocchio: The visitor is a return visitor..... The visitor is a first time visitor....

Thanks for all of those possibilities. I need to look at these different populations. I can easily separate them into... A) visitors who type in the domain or arrive by bookmarks..... B) visitors who arrive by search.... C) Visitors who arrive by links on other sites. I can assume that most of "A" are return visitors. Whatever content they consume might be what has brought them back to the site.

Respree: One thing I found helpful was to have a little script which throws all of your search terms into a simple flat file. You can then import that into a database program (like Access) to find out the number of occurences. Maybe take a look a some of the high volume search phrases and then enter then into your search engine.

Clicktracks will show me what is being searched.... but I need to do what you say and see what my visitors get back for some of these searches. Most will lead to category pages, I think, and I can then track the actions of people who enter a category page from the search page. Thanks This can also inform content development.

eKstreme: 1. Put a "recent searches" list on your home page and monitor its usage.

2. Put a tag cloud of the most popular topics on your site. Hack the CMS to generate that.

Thanks! Great ideas... I might be sneaky and only include the recent searches that work exceptionally well. lol

Autocrat:.... cookies.... condition codes.....

This sounds very valuable... I need to learn how to do these things. Great ideas. Thanks

Wit: Umm, Egol, do these search phrases have lots of HTML tags in them, perchance?

I don't understand, Wit. Are you asking if code is being entered into the search box?

[Wit]

Yep.

[EGOL]

I have not seen any code in clicktracks... but have not looked down at the single query level.

My data on click percentages is coming from crazyegg so those are genuine clicks and not robots. (I don't think that crazyegg sees robot activity).