9 replies to this topic

[identry]

My company recently replaced the Web site of a non-profit organization. The old Web site had been online for many years, but was old-fashioned, disorganized, and difficult to maintain.

The new site has pretty much the same content. It's just reorganized, with a CSS design and a consistent navigation system. Nearly everyone agrees its a big improvement.

Unfortunately, the previous Web master is claiming that traffic to the site dropped by 90% on the very day that we switched over from the old site to the new site. This has stirred up a lot of trouble for us.

We installed Urchin Web measurement software on the site before switching over, so we have visitor data from the first day of operation (about 2 months ago) until now.

As you might expect, the number of visits vary from day to day, but on average, the visitor and page view graphs are flat from the first day until now. That is, traffic has not gone up or down since switching over.

What this Web guy is saying is that the day before we switched over the site, the traffic on the old site was 10 times higher than we measured afterwards. In other words, that switching from the old site to the new site instantly cut traffic by 90%.

The old site was on a different server. All we did to switch over was to switch the domain name DNS entries.

My question: Is this even remotely possible? If so, what could cause such a drop-off?

Any thoughts, much appreciated.

Brgds: John

[OldWelshGuy]

Yep this can happen. Thre is a set procedure for transferring a site, and switching off the server once DNS has propogated is not the way to do it. The search engines take a while to catch up you see, they use IP addresses to spider and do their stuff, it takes longer for them to catch up.

What you should do is leave the old site up on the old server, and set it up on the new one, you ONLY take it down from the old server when the spiders no longer visit the old server and not a moment before.

Some of the things you said concern me. like

1.

The old Web site had been online for many years, but was old-fashioned, disorganized, and difficult to maintain.

2.

The new site has pretty much the same content. It's just reorganized,

3.

with a CSS design and a consistent navigation system

4.

My company recently replaced the Web site of a non-profit organization

1. So you took down a long standing site organised it and put in place some sort of CMS (presumably). What about the urls? did they change? if so how did you handle re-direction?

2. Same as above "re-organisation planning is CRUCIAL to a redesign, you should always try to keep the same urls if at all possible, and simply add the content and try to work with it. If this is not possible then you must sort out the redirection to a tee.

3. Changing the navigation can have a massive effect on a site as this is how the search engines work out hierachy etc

4. Non profits attract a lot of backlinks to deep content, if you have changed the urls and dumped those pages then you will have thrown away LOTS of backlinks. Add to this all those bookmarked and forum linked pages that are now going nowhere, and you have just lost yourself a whole lot of traffic.


So I guess the simple answer to your question is YES. BIG AND FAT

[Jill]

What this Web guy is saying is that the day before we switched over the site, the traffic on the old site was 10 times higher than we measured afterwards. In other words, that switching from the old site to the new site instantly cut traffic by 90%.

It is not only possible, it happens to site in this situation all the time.

As soon as you go removing all the previously indexed URLs for ones that the engines don't know exist, you will definitely lose traffic. Especially if most of their traffic was from search engines.

Better get a PPC campaign going asap, and also start whipping together some 301-redirects from your old URLs to the new ones.

This is just so common, it's amazing that so many webmasters never think of this stuff in advance!

[OldWelshGuy]

This is just so common, it's amazing that so many webmasters never think of this stuff in advance!

You get my vote with that one Jill, I am stunned that many 'expert companies' do not know what is in effect a minimum understanding of the search engine workings.

[jehochman]

Web developers sometimes consult an SEO before overhauling a site to avoid these kinds of problems. I suggest you get professional help ASAP before this redesign causes long term damage.

Looks good and works good are two completely different things. For instance, if the new menus use Javascript or Flash, they may look wonderful, but the search engines may not be able to index the site.

[identry]

Or, it could be that the old web guy was simply un-informed...

This morning, I managed to get my hands on the old log files and ran them through Urchin, so I could compare apples with apples.

Since the old site didn't have any visitor tracking, it's difficult to say anything precise about visitors. However, comparing the month of May with the Month of June:

Page views went from 18,797 to 20,143
Pages views per session went from 2.3 to 4.2
Length of session went from 3:07 minutes to 4:20 minutes

The primary conversion event on this site is for visitors to search the database of members. The number of searches went from 1516 to 1419, a statistically meaningless change, I think. Perhaps because visitors were finally able to find content that had previously been findable only by digging through layers of pages and links.

This data seems to confirm my belief that simply switching the URL will NOT instantly (I mean at the moment of the switch) impact traffic. How could it? If someone does a search on google and clicks on the link, and assuming that the link isn't to a dead page (there are no dead pages... all redirect correctly.) then the link still works.

The same is true with inbound links from other web sites, and bookmarks. All of these will still find the website, so why would people suddenly stopp clicking these links?

I could believe that search engine hits could drop off *after* the search engine has spidered the new site, but until that happens, I don't believe that anything could cause an sudden, drastic change in visitors.

And the statistics, when read in a consistent manner, confirm that no drastic change occured.

I do see a dip in search engine traffic a couple of weeks after switch over, but that traffic recovered fairly quickly... within a few days. So I guess the guys at google, etc., managed to figure out the new site structure, some how.

In the future, I will be very careful to measure the old site before switching over, both as a baseline for improvement and to avoid absurd claims from the jilted web guy.

-- John

[torka]

If the old web guy wasn't using Urchin, it could also be at least partially attributable to a difference in measuring tools. The most important thing to keep in mind with any analytics solution is that none of them are 100% "accurate." All of them have to make certain assumptions under different conditions, and if two of them don't happen to have made the same assumption, they may report very different numbers.

It could also have to do with a (very common) confusion between "hits" and "visitors." If the old web guy didn't know/understand the difference, he could have been reporting "hits," which were being compared with "page views" or "visits" on your side of the reporting divide.

But, yes, your approach going forward is definitely the one I'd recommend. Run the logs through your chosen analytics package to establish a consistent baseline before making the change. As you say, you're then sure of comparing apples to apples.

--Torka

[OldWelshGuy]

Well done John, had you made it more clear in your opening post that you put in place all the proper redirects with belt and bracers, then the replies would hae been totally different. ALL of my replies were based on no redirects being used (as you did not mention this in your post )

This data seems to confirm my belief that simply switching the URL will NOT instantly (I mean at the moment of the switch) impact traffic. How could it? If someone does a search on google and clicks on the link, and assuming that the link isn't to a dead page (there are no dead pages... all redirect correctly.)

That is the difference right there you see. Your question was 'is it possible' our reply was YES, had you stated the above, then the answer would have been a big fat NO

[Jill]

Yep, ditto what OWG said!

[identry]

Sorry... at the time of my first post I didn't have the log files... I really wasn't too worried when I posted the question, but your answers freaked me out, so I demanded access to the original logs.

One of the absurdities which had me just slightly suspicious was a table created by the old guy with columns like:

Date: June 2005
Hits: 120K
Pageviews: 20K
Vistors: 27K

H'mmm...

Anyway, been reading your newsletter for years, Jill. Keep up the good work.

-- John