Can't see the wood for the trees?

Sometimes I feel as though customers allow their attention to detail to run away with them too much and they can't see the wood for the trees. There's a coupld of customers' sites that are built and ready to go live, but in each case something is stopping them that's just too much attention to detail.

In the first case, the site is wonderful, works well and has had the stock loaded to it. It's published and ready to sell, apart from one missing essential - a home page. The home page isn't there - so anyone typing in the URL assumes there's not a site there. It's a shame when so much effort has gone into the site and the URL is being advertised.

Why isn't the home page 'ready'. Well, she gave us a logo to use and basically just want's this large graphic banged slap in the middle of the screen. Without checking, it's 500 pixels wide by probably 400 high. So it fills up most of the screen. But she want's it a little bit wider. Not much, just a little. So she's gone back to the original graphic artist. When she extended it and I published it, the customer didn't like the feel. So it's back to the drawing board. All for the sake of a few pixels. The graphic looks fine and I'd be happy to be a customer of the site.

In the second instance it's just nit-picking over fine details on the screen - those links in the top right should be the same size as those on the left was one of them. They were 8px and 9px Arial font respectively. Then some pictures needed 'balancing' more accurately - they were moved apart by 2 to 3 pixels. And the list of changes goes on.

Why? OK, if they make the site look horrible then yes, get them sorted. But most customers aren't going to notice these problems, and he has the ability to change the pictures anyway - the sizing will affect the spacing.

What is wrong with this? Well, apart from wasting time in development, the sites are on hold, not earning the customers' the income they should be. Another recent customer picked over his site and finally agreed it was finished and then started receiving huge amounts of orders each week. That's that many lost each of the previous weeks.

Does it matter if these details aren't right? I think not. I like to review the style of the site anyway once it's live and make changes if required then. As long as the site looks good and gives the site's customer the feeling that the shop is professional and not going to run off with their money, what does it matter if links are 8px or 9px? If they can be read, they can be read. If the site looks good enough to convince people to buy the products, they will buy them.

Geocoding ain't so easy!

My redesign of Cottages-4-Holidays hit a stumbling block last night when I tested how accurate the postcode to longitude & latitude file worked. I downloaded the file a few weeks ago, when I first put my ideas together, and was really excited about what I'd found.

But when I tested it, none of the properties I expected to see where there. I'd test loaded a popular tourist attraction and looked at when properties I was claiming were close by. I checked these against the list that mentioned the place in the advert and they weren't included. Then I looked at where the properties were - and they were well off!

It doesn't affect the main build of the site, I can continue there. But it does affect how quickly I can load up tourist information. The problem is that all longitudes and latitudes were being rounded down to the previous whole number. So -1.988 etc become -1. Given at the same time 52.898 or whatever became 52, that was appearing as a fiar movement across the country.

The cottages themselves have geocoding supplied by the merchant, so I will trust them. But my own geocoding needs looking at. I have basically 2 choices:

First - the quickest choice is to again guess the geocoding based on the postcode - look for cottages on the database with similar postcodes and assume the same geocoding holds true. Probably a lot less effort, but not as accurate and what when there's no property with a matching postcode?

Second - every time I find somewhere to add, which will be quite a lot at first, I need to use a tool to geocode the postcode. Much more accurate, but it's adding time to the process and that will mean that in the end I get fed up quicker!

I suppose if the stats show that the site is starting to work again then I'll find the time and energy to put into lots of properly geocoded tourist attractions. But I was going to write an admin function to load new ones and get my wife to give me a hand! She'll just have to learn how to geocode as well!

Cracking the XML

After a lot of struggling I finally got the XML feed working and the first 2 directory pages, linking to 9 merchant pages are up! I started on the 2 smallest categories I could find and even they took a while - and I'm not even 10% of the way through the list of merchants! It should get quicker as I work through them, I was still ironing out problems and smoothing the system into place at first.

So I've created an XML based feed for Mortgage Providers and Car Breakdown Providers. I suppose that by adding 1 new section a week, or maybe say a handful of merchants each week, then that's probably best as far as the search engines are concerned.

If I go full out to get all of the merchants up it would be a good sized directory structure, but there might be more appeal to the search engines if I can build it steadily. That way, every time they return to the site there are new pages to uncover. It's giving a more maintained look to the site. And along with the merchants updating their XML feeds, this will provide a series of pages on which there are regular updates.

Hopefully, this area will be wonderful bait to the search engines and will start driving in more traffic to the site, and maybe even a few more affiliate commissions. It wouldn't take many affiliate commissions for me to be really happy. Some of them pay really well and just 1 payment would out a huge smile on my face!

I did decide last week, aside from this area of work, that to get traffic back to what it was on CompareMortgageRates I probably needed to start adding new content pages weekly. Well, whilst I work out what I'm going to add this is a good starting point! I'm also trying to add 'news' to content pages (as well as the home page) to also keep these pages fresh. Since the site made it's drop through the search engines I've been trying my best to recover it. One day I will, but the traffic has dropped further over the last week, even though it's position on Google searches remained constant. Presumably other search engines have now also lowered it!

XML Feeds - Made to make you work?

I decided last night to have a play with the OMG XML feed for their editiorials, intending to use it to display merchant information on my mortgage rates website. Basically, starting a financial directory that I can later add to. In the day I'd noticed another site that ranked quite well doing this, but manually updating the displayed information (the offer details were out of date).

Now OMG provide their editorials in 3 formats - javascripts, which I usually use, I-Frame, which I've never tried, and XML, which I'd not tried before from them. I've used XML from other providers plenty of times - other affiliate schemes, news readers etc. So I've got plenty of the basic code about.

The problem with Javascripts and I-Frames is that it doesn't add anything to your webpage as far as the search engines are concerned - to my knowledge, they just ignore these parts of the code. At most, they will actually follow the I-Frame (I have seen that in some of the sites I've built), but they the benefits of the content lay with the provider. There's not lots of pages of text that I can create and add further bits to.

So taking the XML feed in real time seems a good idea. Why not just copy the text?Well, as I mentioned above the site I saw yesterday did this and said that the insurance offer was a 15% discount, whereas it is now 20%. OMG don't like affiliates taking the text alone as out of date offers and rates are displayed. And I don't want to create a large nightmare for myself of continually having to update text - it's bad enough when they don't update the dynamically served content and email me to complain (you know who you are if you are reading this!!!).

So last night I started to build into my standard code the OMG feed. It looked as though it would be easy - not exaclty a hard layout to use. But I was playing with it for hours as it just wasn't working. 'ROOT' was appearing at the end of the text - it's actually from the closing tag (<\ROOT>) and control characters were appearing where they shouldn't be - something like &#xA; instead of carriage returns. I thought it should be simple, change &#xA; to <br>. But it just didn't work. I could remove the & on it's own and the #, but the string just wouldn't go.

Then when I looked at the output in notepad it appeared that the text was full of non-printable characters. Between every displayed character appeared a space in notepad, which didn't appear on the screen. Not knowing what these were, made it very difficult to remove them. I suppose some sort of regular expression could have done it - just thought of that now!

What their purpose is and whether it's something I was doing wrong, I just don't know. But something wasn't right. It could be that they are there intentionally to make sure that the text isn't read by the search engines - either to stop them caching text which goes out of date or to prevent problems with Google's duplicate content filter. Either way, it was a pain I was trying to sort for about 4 hours. I had hoped to get the feed going in 30 minutes and spend the rest of the time getting most of the pages up. It wasn't to be.

Would appreciate any more thoughts on this problem and hearing what others have done.

Which Webpages Get Page Rank

Back to looking at which pages get page rank and which don't inherit anything.

Recently I took on a new customer with an existing site. He was a bit upset that his current webdesigner was going to charge £60 + VAT just to renew the URL (the guy is using the same registrars as me and they charge under £8 plus VAT). He was also shifting his concentration slightly away from scooters and more to motorbikes, so wanted a copy of the website under a different name.

It's an OK looking website so it was just hosting changes, plus I gave him a content management system so that he could change the bikes he had on display - with his previous designer he had to send him the details and once every few months the bikes would be changed. That was another bug-bear!

So I took over the website, created new pages to display the bikes (PHP rather than HTML to access my database) and created a copy of the site, with a different sort order - let's avoid that duplicate content filter.

He's just asked me to make a slight change to the sites, so I was looking around the pages. Most of the pages, including the new ones, are now in at PR0. In fact, both sites have the same distribution of Page Rank.

On both sites the home page is PR0 - not something to be proud of, but then it's not an SEO optimised site - it's a contact point for his magazine adverts.

On both sites all of the bike details pages are also PR0.

The enquiry form, with very little content (just the field names) is PR0 on both sites.

The contact form on both sites, with matching address, phone numbers and email address (word for word the same - both display the one email) are both PR0 - so much for the duplicate content filter...

But the 'company' page - with details about the company and 125 words in the text as grey barred on both sites. Both pages are cached, so it's not that the search engines haven't found them.

The only major difference, and this is because it's a site I've inherited rather than designed, is that the company page is reached only by an image map link, whereas there are text links back to the bike pages, the home page and the enquiry form. There's not even alts / titles within the image map - something I would have done if I'd written the site myself / been asked to optimise it.

But this alone is not stopping the inheritance of page rank as the contact page is also only available through the same image map. And that's got PR0. Depending on how you count it, it's only got around 20 words of text on that page.

So, the company page isn't grey barred (on both versions) because it's only linked through an image map and it's also not grey barred because it's not got enough content - the contact page has less. So what can it be?

The smallest bikes page has around 250 words on it, along with pictures of 4 bikes. There's no links to other pages or anything, you have to phone for details.

So I think it's one, or a combination of:

- Google doesn't like the duplicated information, but is happy when the duplicate information is contact details.

- Google doesn't like the content.

- Google is cleverly detecting which pages could be of interest to people.

I suppose it is the pages that are most likely to be of use to people searching the internet that have page ranks. People aren't likely to be searching for the company history, but might want to contact them after seeing a magazine advert or might be searching for an offer on a bike.

I think a few more investigations are required to see the effect of page rank on other sites I manage.

What does the redesign include?

Well, for a start the redesigned site is still under development so it's hidden away. So I can't show you it (could get confusing if search engines got hold of a development version!). But here's the features I'm including and why. Then, unless I make excellent progress and get the new pages ready to go live, I'll leave this theme for a bit and look at what I was on before I distracted myself!

First, I'm going from building it on my pc to loading to a database - this means the process is quicker and can conceivably be achieved in around 10 minutes' work - to be proven when I write it!

Second, I'll now be able to identify new properties and highlight them on the home page and other main pages - getting cottages listed quicker.

Third, I'm building (this is going to be a long ongoing process) a database of UK tourist attractions. Each property & town will link to the nearest ones I find. This adds unique content, and people might find the site searching for these attractions.

Next, when the owner has mentioned a tourist attraction in their text that I have a page for, I'll change the description to link to it. Makes the internal linking fuller.

Next, when the owner mentions a town in a description, I'll link to that. Makes the internal linking fuller.

Also, I'll display the nearest properties to the one shown, just in case it's not suitable - it also helps add more content to the page.

Possibly I'll also take a feed of hotels I have access to and include nearest hotels on the page - again, more content and possibly more commissions.

I'm also adding a Google map showing the property - adding more value to the customer. But the map doesn't seem to want to work at present... Works fine, until I try to include it in the code...

I'm also monitoring which pages are hit and linking to them as the most popular etc from the home page - helping customers find the most popular properties; telling me which are the most popular properties and getting the search engines faster links to these properties.

I'm sure there's more, but that gives a flavour of what the redesign is all about.

Benchmarking The Redesign

How will I know if the redesign of the website is actually of benefit? What would be a benefit - what makes it worth my time redesigning a site that has basically flopped after a once proud time?

In short - income! This is derived from 2 factors - search engine visitors arriving on the site who then either book or click Google adverts.

Now, I've already said that the success of the site varies month on month. January / February are usually the best times of the year for the site. So what's a good way to benchmark the site?

I could look at the current position and compare that, but I wouldn't know if changes, or more to the point the percentage of changes, are being derived from my work.

So the best way forward is to just update one of the 2 sites - the least well performing one, wait for the search engines to pick up on the changes and hope that I see massive swings.

Last month, the second site got just 4,942 page impressions and the first one a mere 992 page impressions. I used to get that per day! The income per site was $170 & $31 from Adsense (I can't mention click through rates / cost per click - Google policies). And between them, they managed 2 failed bookings and 5 confirmed bookings, valued at £90.48 commission for me. So around £200 income in the month. Not fantastic, but it's something.

So that's where we start from. Hopefully in a few weeks I'll start to see the ratio of traffic change from it's current 5:1 to (maybe) the other way around and the incomes improve accordingly.

Tomorrow, what I am doing to achieve this.

Why To Redesign When Website Traffic Tumbles

So what is the driving force behind me wanting to redesign a website that used to work perfectly well? Simple - it's not working so well now, and I 'blame' the Google duplicate content filter for its demise.

Blame might not be the right word, but you get the drift. So what happened?

Well the first holiday cottages website got loads of traffic, more than the second holiday cottages website one by about 7:1, if my memory is correct. That's when they were both at their peak.

Then they both dropped off and not much happened. With the second being newer, having less traffic and therefore being a lesser risk if anything went wrong, I decided to convert it from html to php and make a few changes along the way. Basically, the intention was to make it quicker to upload (using include files for the standard code which meant 4,000 smaller files to ftp...). But at the same time I added a few small tricks that allowed me to change the descriptions and re-write a lot of them.

I put this in, sat back and waited. Along came the search engines and after a few weeks, onces all of the pages were cached, the traffic increased. In fact last month it was 5 times the first site's traffic, although still not enough!

Confident that was the problem with the second, I did attempt a minor rewrite, but never having the time it was half put in and probably made the site worse.

So a few weeks ago I started again (I've frequently pick up the site after over a week of not working on it...). My idea is to produce a site that doesn't just show the affiliate content but also has plenty of unique content and is internally linked in such a way as to provide a network of information. Hopefully, something the search engines will like and send visitors to, and something visitors will be able to make use of and want to book through...

Why Redesign A Website?

So I explained yesterday that I have 2 cottages websites, built slightly differently from the same affiliate data feed. The original was getting far more traffic than the second, because of the keywords (presumably) that each targeted. So what happened?

At first, I expected 700 page impressions to one confirmed booking! Yikes, that's a lot, but that's based on the first year or so. Some months are better, some worse. So why is this?

Well, for a start it's difficult to track what's happening as the affiliate scheme doesn't provide tracking so I don't know which of the 2 sites gets each booking. Both are different styles, so one style could work well, the other could be failing. And the success changes throughout the year - as availability changes.

Now a lot of these people will be looking and then realising they are seeing cottages they have looked at elsewhere (other affiliates) or they could be looking and checking availability and finding the accommodation booked. It is frustrating that although holiday makers are wanting to book from the end of the year, the availability doesn't tend to go onto the database until the start of the year. So there's lots of lost enquiries there.

There's also the worry that with that low conversion rate people are seeing the site and thinking 'yuk' - quite possible in the original site, but the second looks much better. And I suppose that there is also the confidence factor - do you really trust the site enough to book?

But the main reason I am redesigning my site is that the traffic has fallen off dramatically recently. I'm certain I know why - more tomorrow...

Redesigning a Website

I thought I'd take a break from the website marketing and talk through the process I'm currently going through with my Holiday Cottages website. It's my own site, that's been running for almost 3 years, and I'm about to put in a totally new version of the site. It seems a shame not to share what I'm going through and record the frustrations and hopefully successes of the site.

But what is the site and why am I redesigning it?

Well, it started off around 3 years ago - just over a year after I set up my webdesign business. Things were quiet and I was looking for other avenues to supplement the income. It had never been my intention to completely run the business as webdesign - I don't like the eggs in one basket situation.

I'd fallen quickly into affiliate marketing - starting off with the Mortgage Rates website. Now there's another site in need of plenty of TLC! That started as a venture between me and a mortgage broker and I'd accidentally discovered affiliate selling and started trialling banners to supplement the income. 4 months later I discovered Adsense and tried that (very successfully) as well. At the time the site was getting a lot of money spent on it in Adwords, so the income was well received.

Then one day I discovered an affiliate scheme for holiday cottages and decided to give it a go. I was able to take a feed of all of their properties, create almost 4,000 pages of website using Perl and then FTP the site to my hosts.

But it takes ages to load that - every month!

After about 4 months the site had taken off, it was getting reasonable traffic (for a site that only needed me to initiate the monthly rebuild) and was earning £400 per month plus on commissions plus Adsense income.

So I built a second site, hoping that would be as successful, but it got about 15% of the traffic (it was written differently to cover different keywords). But still, the bookings went well for a few months.

Then the traffic dropped off, the original site became grey barred and I was too busy with customer's sites to do much else. Total income between the sites is around £200 per month, which is still a good return for the monthly effort (or lack of effort).

But was was wrong, what could I do and what have I tried? Read on...