Ecommerce - are you ready for the next stage?

movietub

Free Member
Nov 6, 2008
4,858
1,103
I've been off the boards for quite a while now, working furiiously on the version 2 of our own website. The current site has been more succesful than we ever predicted, but I'm aware ecommerce is shifting up a gear. I thought that I would impart some of what I have researched as a *hopefully* useful re-entry to the UKBF community :) I have also spent a lot of time reflecting on what is the likely direction of ecommerce in the near future.

So in the first 10 years of the century, online shopping 'happened'.
Yes you can trace the genesis of online shopping way back, but the turn of the century was the time the whole of the west seemed to be focussed on technology. Minds opened up, the digital revelotution was started propper. And like mobile phones, broadband and satnav, online shopping was about to become common cause for discussion, not just the territory of the geek.

And how did businesses respond?
Oddly, or perhaps not, as many independant start-ups jumped on the ecommerce bandwagon, as did the established high street names. In many cases names that had conquered the traditional retail world took longer to get a website sorted out than some of the more enterprising individuals.

And how did SMEs in particular respond?
Naturally we fit into the most enterprising business group of all. Most of us on these boards are small enough to flex and adapt very quickly (not that all of us do so willingly). Soon there was a multitude of quick and easy solutions to get started selling online. Certainly for the last five years the common question has been not 'how do I sell online' but 'What do I sell online?' The masses have heard that selling online makes easy money, and the only barrier to entry was finding something to flog for more than you can buy it for.

So why will it become harder to start up and sell online?
Because online shopping is now so commonplace, certain unwritten rules are emerging that need to be addressed. This is of course in addition to the written rules which also often get overlooked...!

Delivery times
Expectations are shifting fast. Speed of delivery needs addressing, which means holding stock. Customers increasingly consider guarenteed quick delivery equally important as price. The requirement to hold stock is in many industies no longer optional, its what the shoppers demand. This in turn requires a site that is able to display stock levels, which relies on a back-end system to track stock. Entry to the market is made much more difficult for the basic 'site in a box' users. It is also making life harder for current traders that are limited by their chosen cart solution or their own time/ability/willingness to move up a gear.

Navigation
Now that we have 10 years of online shopping under our belts we all have higher expectations - we have seen how easy to use some sites can be, and how awkward many others are. Now is the time then that the big players are leading the way, storming ahead in fact. visit http://www.diy.com and have a browse round. This is an example of a cutting edge site that employs the following:

Mega MenuThe top menu, hover over it and you will see the 'mega menu' theory in practice. This is being adopted by an increasing number of big retailers and its easy to see why. Unlike traditional dynamic menus which either cascade several times, or simply limit the number of options available, the mega menu offers at a glance organised access to pretty much any page on the site. It's the culmination of just about every other method of creating a dynamic menu to date. They can be designed using pure CSS, great for SEO and as each drop down can contain any content required. Images, flash, links and other dynamic objects. Its essentially possible to offer click free access to user relevant content that used to require its own page. Many articles attribute the birth of the website 'mega menu' to MS use of similar dynamic content containing drop downs in their ribbon menu system.

The good news is that as a pure CSS solution anyone with access to their sites CSS has the ability to implement one. Many sellers feel limited as they cannot write or use JS on their site, well this a trend that is gathering pace at the top end, it makes sense and you can get on board with it now.

Intelligent Search
Despite us all benefitting from ever more intelligent search engine algorythems, the majority of ecomerce sites still use very basic search systems. I.e., most sites require the user to input the exact string of letters that are found in the product name. On our current site, searching for 'fish nets' will not bring up a product called 'fish net'. This is not always easy to solve but always possible - and its crucial. You absolutely cannot sell something that a shopper cannot find. Google custome site search is usable in most cases, and for a reasonable cost, can be used as a fully integrated or customised part of your site. Other solutions are available of course. In fact, we nearly moved of to Magento because of the progress they (and the various extentions) have made in this area alone.

Standardised layout
Now true standards exist of course, but there are common layout trends which are emerging as popular. Most common desktop software, across various platforms as well, looks vaguely similar in layout. This is not due to lack of imagination. It's due to the fact that the users are used to looking in a prticular area for a prticular function, its about making the user experience easy and comfortable. Be different where you can, that is the joy of the internet, but pay attention to developing trends. A common example is forgetting to place contact info and links to legal blurb etc at the bottom of each page.

Learn basic HTML, CSS and image creation.
In the same way that a tradional shop owner does not need to know how to build a shop, you do not need to know how to build an entire site. But you do need to know how to dress and organise one. Any traditional shop owner can stick an attractive SALE pos in the window, so it follows that anyone serious about selling online needs the basic knowledge to achieve the same attractive presentational tricks online. The most useful thing a web designer will ever be able to do for you is help you understand what you really need, to achieve the functions you want, and build that framework. Once you have that, you need to take over on the day to day stuff - and its stuff that needs doing to keep a site fresh and current.

Do SEO - the really easy SEO that requires no understanding of SEO...
Content really is king online. So many independant sellers simply copy and paste supplier blurb for thier products. Search engines see this as one of the lowest form of page content possible. There are many things a professional can do to help search engines crawl your site easier, and to develop important keyword strength. But ultimatley nothing will bring you a greater return on time/investment than well written, unique, useful content. Always write in a Q&A style where possible, most searches start with either a question or an answer, if your text matches either closely then you are well up in the rankings. Again, basic SEO is something that any of us can research and understand.

Get professional
Sharpen up the copy, tone and layout you use on the site. 5 years ago shoppers were an experimental bunch, early adopters willling to trust most sites so long as they had what they wanted at the right price. Now shoppers are smart and increasingly educated regards how to identify warning signs of a trader that is either amatuer, or likely to be unreliable. This ranges from the obvious crime, no phone number/address... To the more subtle. Is your menu system a mess? DO colours clash? Does your site look like its matured and developed and updated well? Or does it look like it could have been built from a basic template just 2 days ago? Is your wording concise and clear? Do you use to many'!'s!?!??!?!?

Presentation is becoming so important. Because anyone can have a website, of any size, for free. Physical shops cost money, having the biggest or best presented shop on the high street costs millions. Online any site can be as big as you want, or presented however you want. So present well and strive to judge your efforts against those of your competitors.

Growth
Growth is important too at the moment. Economy of scale and buying power are becoming vital to maintain the competitive price and high levels of service shoppers now start to expect, often demand, online. You cannot enter a marketplace and grow if the effort is only causual. A website is a business like any other and must stand strong with one eye on the future. Expect the quantity and variety of ecommerce websites in any one sector to diminish gradually, as on the high street before. Ultimatley those that move with the times and grow their own brand strength and values will be most secure. Secotrs already exist which have effectively been narrowed down to just a few major players, making entry now almost impossible without big and long term investment.

Finally - just take ecommerce seriously.
The above points are a handful of what I could have gone into. The undertone is that over the last ten years the party has been open to all-comers. But now, to move forwards or enter the market as a serious business venture, you need to be very serious about what really makes shoppers tick. The 'easy - basic' online shops out there will increasingly feel the pain of limited functionality, flexibility and power. And you cannot escape those bounds without developing your own knowledge and skills at the same pace as the marketplace develops.

David
 
Funny you should mention diy.com. I was looking for roofing felt yesterday on there. They have it in the store but if you search for 'roofing felt' on diy.com it takes you directly to a rabbit hutch and if you search for 'felt' there are three pages of felt things, none of which are roofing felt. I ended up just phoning them and they have several types and lengths available. Perhaps their menu is good but their site is very lacking.

Nice article though. :)
 
Upvote 0
Some interesting opinions, thanks.

I'd just like to pick up on one point regarding the the diy.com site and "mega menus":
Mega MenuThe top menu, hover over it and you will see the 'mega menu' theory in practice. This is being adopted by an increasing number of big retailers and its easy to see why. Unlike traditional dynamic menus which either cascade several times, or simply limit the number of options available, the mega menu offers at a glance organised access to pretty much any page on the site.
As with most things, there is always some impact with implementing new stuff. Some new stuff might appear to have a lot of advantages, but there may be detrimental affects too, so be aware of these.

Mega menus can sometimes push the amount of links on your home page to a very high number, and there may be some detrimental SEO affect because of this. The diy.com home page now has just over 300 links!

I'd like to pick up on the phrase "hover over it" too. New web browsing devices such as the iPhone, the iPad, other touch phones and other tablet and touch devices coming soon, which are all touch driven, do not have the concept of the mouse "hover over" (the equivalent of moving a mouse without any buttons pressed). So don't rely on hover, and if something is important enough to popup when you hover over, try to find an alternative way of showing that for non-hover devices. At least with diy.com they do allow the hover popups to appear if the menus are clicked on (touched) too.
 
Upvote 0
Intelligent Search
Despite us all benefitting from ever more intelligent search engine algorythems, the majority of ecomerce sites still use very basic search systems. I.e., most sites require the user to input the exact string of letters that are found in the product name. On our current site, searching for 'fish nets' will not bring up a product called 'fish net'. This is not always easy to solve but always possible - and its crucial. You absolutely cannot sell something that a shopper cannot find. Google custome site search is usable in most cases, and for a reasonable cost, can be used as a fully integrated or customised part of your site. Other solutions are available of course. In fact, we nearly moved of to Magento because of the progress they (and the various extentions) have made in this area alone.
Ecommerce sites having poor search facilities is sometimes a problem associated with the underlying database technology used, and whether that database supports full text search or not.

For instance, ecommerce sites based on MS Access might not have full text search because MS Access does not have full text search (AFAIK), whereas ecommerce sites based on MySQL or MS SQLServer should have full text search.

Some ecommerce systems which do use MySQL, might still not be configured correctly. For instance, by default osCommerce is only configured to search titles not descriptions, an osCommerce configuration/coding issue.

There may also be some database configuration issues, e.g. how small a word is included in the full text index, this is often set to 4 character words or above only, which is no good if you are expecting your customers to search for "net".
 
Last edited:
Upvote 0
Some interesting opinions, thanks.

I'd just like to pick up on one point regarding the the diy.com site and "mega menus":

As with most things, there is always some impact with implementing new stuff. Some new stuff might appear to have a lot of advantages, but there may be detrimental affects too, so be aware of these.

Thats my general take on these things as well. I feel differently about the mega menus because it is possible to generate in pure html/css, so no script issues. This was never possible with multiple layer menus because using CSS only you cannot set delays. It's near impossible to hover and open up several levels of a menu without slipping off and and the whole lot dissapearing before making your selection!

Mega menus can sometimes push the amount of links on your home page to a very high number, and there may be some detrimental SEO affect because of this. The diy.com home page now has just over 300 links!

Yes, but this is a real pros vs cons area. Hundred of links do not do much to support each page having dedicated relevant data, but on the other hand if all links lead to a page which does balance that with quality content then it should not be an issue. There is also the need to balance usability against pure SEO focus. Who cares if perfect SEO gets an extra 20% traffic per day if people then leave because the nav is limited or congested? I'm pro navigation ahead of SEO currently, simply because I know navigation is useful, and no-one really knows how Google etc react to a certain number of links. I would like to think that as all the big name stores are adding so mnay links to each page however, that there is some smart money riding on it somewhere! I'm also fairly sure that google are not blind to the fact that there is a requirement for menue systems on large sites that will need to contain many hundred links. It would seem sensible if they judged the number of links in comparison to the site size, but who knows?

I'd like to pick up on the phrase "hover over it" too. New web browsing devices such as the iPhone, the iPad, other touch phones and other tablet and touch devices coming soon, which are all touch driven, do not have the concept of the mouse "hover over" (the equivalent of moving a mouse without any buttons pressed). So don't rely on hover, and if something is important enough to popup when you hover over, try to find an alternative way of showing that for non-hover devices. At least with diy.com they do allow the hover popups to appear if the menus are clicked on (touched) too.

Our upcoming solution, as does B&Q's gets around this simply by making the hover area clickable as well. An iphone user would never know about the hover function, but on clicking the button would get taked to the same options on a static catagory page anyway.

But your right, mobile devices will need more consideration in the not to distant future. Most touchscreen devices currently are rarely used for shopping, people tend to shop online when they are not doing anything else, i.e., not on the move. The introduction of tablets replacing laptops etc could mean that more and more touchscreens will be used for shopping and other dynamic website interactions.
 
Upvote 0
Ecommerce sites having poor search facilities is sometimes a problem associated with the underlying database technology used, and whether that database supports full text search or not.

For instance, ecommerce sites based on MS Access might not have full text search because MS Access does not have full text search (AFAIK), whereas ecommerce sites based on MySQL or MS SQLServer should have full text search.

Some ecommerce systems which do use MySQL, might still not be configured correctly. For instance, by default osCommerce is only configured to search titles not descriptions, an osCommerce configuration/coding issue.

There may also be some database configuration issues, e.g. how small a word is included in the full text index, this is often set to 4 character words or above only, which is no good if you are expecting your customers to search for "net".

It sometimes feels like package cart providers should try harder. It's not really fair to sell a package as a having 'a search function' if it not actually that useful. In most cases it can be worked around, for example Google site search can be seamlessly integrated, but I'm at a loss as to why more intelligent search functions arn't applied as standard. However the db is constructed, there is no reason why it is not possible to interogate the db in a more intelligent fashion. Or at least allow tagging of search terms to pages - many packages can't even do that.
 
Upvote 0
Mega menus can sometimes push the amount of links on your home page to a very high number, and there may be some detrimental SEO affect because of this. The diy.com home page now has just over 300 links!
Yes, but this is a real pros vs cons area. Hundred of links do not do much to support each page having dedicated relevant data, but on the other hand if all links lead to a page which does balance that with quality content then it should not be an issue. There is also the need to balance usability against pure SEO focus. Who cares if perfect SEO gets an extra 20% traffic per day if people then leave because the nav is limited or congested? I'm pro navigation ahead of SEO currently, simply because I know navigation is useful, and no-one really knows how Google etc react to a certain number of links. I would like to think that as all the big name stores are adding so mnay links to each page however, that there is some smart money riding on it somewhere! I'm also fairly sure that google are not blind to the fact that there is a requirement for menue systems on large sites that will need to contain many hundred links. It would seem sensible if they judged the number of links in comparison to the site size, but who knows?
It's actually fairly simple to come up with a solution which offers the best of both worlds. The same Mega Menu with the same DIV/CSS code, but done in a way that doesn't impact the SEO footprint. Here's how:

1. From an external js file included at the appropriate point, document.write the popup menu divs from javascript code

2. Robots disallow the external .js file - so that search engines ignore this particular piece of javascript execution and do not index what is produced

Disadvantages are that it only works in Javascript enabled browsers, and it might require an extra js file include (unless placed in an existing js file).

Advantages include the idea that you're not really changing much of your existing site from an SEO viewpoint, so this minimises the risk of any SEO detriment. Other advantages include the fact that you're cutting down unnecessary bandwidth, placing quite a bit of menu code in a shared cached area rather than downloading it with each HTML page.

In some ways, just as CSS attempts to separate presentation from content, adding a secondary navigation interface like a "Mega Menu" should be done in a way that separates secondary navigation from content too. In some ways the latter is more important because without doing the latter there is more risk of a detrimental SEO effect.
 
Last edited:
Upvote 0
thanks for this really good post - really useful contribution, in fact I would say this is probably a sticky contender IMO

its nice to read an up to date bit of ecommerce advice
 
  • Like
Reactions: movietub
Upvote 0
thanks for this really good post - really useful contribution, in fact I would say this is probably a sticky contender IMO

its nice to read an up to date bit of ecommerce advice

And the truth is... That post was only a tiny part of what we have had to consider for our site this time round!

If its worthy of a sticky, I would probably request I re-write it a little ;)
 
Upvote 0
It's actually fairly simple to come up with a solution which offers the best of both worlds. The same Mega Menu with the same DIV/CSS code, but done in a way that doesn't impact the SEO footprint. Here's how:

1. From an external js file included at the appropriate point, document.write the popup menu divs from javascript code

2. Robots disallow the external .js file - so that search engines ignore this particular piece of javascript execution and do not index what is produced

Disadvantages are that it only works in Javascript enabled browsers, and it might require an extra js file include (unless placed in an existing js file).

Advantages include the idea that you're not really changing much of your existing site from an SEO viewpoint, so this minimises the risk of any SEO detriment. Other advantages include the fact that you're cutting down unnecessary bandwidth, placing quite a bit of menu code in a shared cached area rather than downloading it with each HTML page.

In some ways, just as CSS attempts to separate presentation from content, adding a secondary navigation interface like a "Mega Menu" should be done in a way that separates secondary navigation from content too. In some ways the latter is more important because without doing the latter there is more risk of a detrimental SEO effect.

For the purposes of this post I went down the pure CSS route, as the majority of hosted solutions do allow clients to write their own CSS. I tried to keep the whole post focussed on what people can actually do.

I have already created our new menu using pure CSS as any JS has to either sit within the body tags in the markup, or be planted in an external JS file on another server. Last time I did that Google Checkout didn't like it.

Also where do you sit on the use of JS for vital, or at least preferable functions? Not everyone has it enabled.
 
Upvote 0
Intelligent Search
Despite us all benefitting from ever more intelligent search engine algorythems, the majority of ecomerce sites still use very basic search systems. I.e., most sites require the user to input the exact string of letters that are found in the product name. On our current site, searching for 'fish nets' will not bring up a product called 'fish net'. This is not always easy to solve but always possible - and its crucial. You absolutely cannot sell something that a shopper cannot find. Google custome site search is usable in most cases, and for a reasonable cost, can be used as a fully integrated or customised part of your site. Other solutions are available of course. In fact, we nearly moved of to Magento because of the progress they (and the various extentions) have made in this area alone.

Search makes a massive difference to sales on a large site. People can't buy something if they can't find it. Not a problem on a site with only a couple of thousand products, but a real problem if you have hundreds of thousands of items.

On our new Kartris software we spent a lot of time on the search because it is an area where our previous classic ASP software (along with most others on the market) was lacking.

Firstly we support SQL server full text search where it is installed, but if its not it implements similar functionality within ASP.NET. This means that you have "relevance matching" with hopefully the most relevant matches top. You don't need to get an exact phrase match, but if you do it will likely rank higher than one that doesn't have this...

The search also adjusts based on the size of the database. You can't search quite as thoroughly through a 1million product database as you would for one with only 10,000 products - it'll be too slow. So our search adapts based on how big the catalogue is, to get an optimal balance between speed and quality of results.

We've recently implemented it on a site with over 1 million products.

www.mtmc.co.uk

You can search very specifically by eg product code (eg try "GT501").

But if you do a more general search, eg you search for "dual action asymmetric blades" you'll get exact matches first, but then below partial matches (as its sorting the results in what it considers the most relevant).

And it all works pretty fast considering its searching over 1 million products, thanks to SQL server full text search.

This greatly increases the ability of customers to find what they are looking for, and that in turn results in significant increase in the size of an average sale (typically around 20%).

Better search earns you more money.
 
Upvote 0
Search makes a massive difference to sales on a large site. People can't buy something if they can't find it. Not a problem on a site with only a couple of thousand products, but a real problem if you have hundreds of thousands of items.

On our new Kartris software we spent a lot of time on the search because it is an area where our previous classic ASP software (along with most others on the market) was lacking.

Good for you! We're commited to our current solution but 2 years ago the above statement would have made me look very seriously at your offering.

It's quite amazing how many of the more popular closed source solutions have such glaring ommisions. There is no reason why even an entry level, low budget, cart solution should not have features which clearly make sense.
 
Upvote 0
mtmc.co.uk

You can search very specifically by eg product code (eg try "GT501").

But if you do a more general search, eg you search for "dual action asymmetric blades" you'll get exact matches first, but then below partial matches (as its sorting the results in what it considers the most relevant).
To test ecommerce search capabilities I sometimes look for phrases in product descriptions, see if the product comes up in the search for those exact words, then play around with the words, typing other combinations that users might type, to see if the product still comes up.

In this case if you search for:

asymmetric blades dual action

the results change, and the original product mentioning all of those words doesn't appear near the top. So this suggests that there are still improvements which can be made in the search relevance algorithms.
 
Upvote 0
To test ecommerce search capabilities I sometimes look for phrases in product descriptions, see if the product comes up in the search for those exact words, then play around with the words, typing other combinations that users might type, to see if the product still comes up.

In this case if you search for:

asymmetric blades dual action

the results change, and the original product mentioning all of those words doesn't appear near the top. So this suggests that there are still improvements which can be made in the search relevance algorithms.

If you change the order of the words then the result will change. The order of words is important. An exact phrase match will be worth more points. If you change the order, you don't have an exact phrase match. This is exactly how relevance search should work.
 
Upvote 0
If you change the order of the words then the result will change. The order of words is important. An exact phrase match will be worth more points. If you change the order, you don't have an exact phrase match. This is exactly how relevance search should work.
What you are are saying about order being important is true.

But in this particular case Google, and users, would disagree with you that the most relevant results are being returned.

If you compare the site's search results for:

asymmetric blades dual action

with google's results, you'll notice that google places different products top, the hedge trimmers which mention all of those words, the same types of products the site places top for the other search: dual action asymmetric blades.
 
Upvote 0
Also where do you sit on the use of JS for vital, or at least preferable functions? Not everyone has it enabled.
I'm a firm believer that you cannot make "everyone" happy when it comes to web design. You have to play the numbers game, check the stats for your particular market, and work from there.

Browsers without Javascript enabled is just one issue. What if cookies aren't enabled too? What if images aren't enabled? What if some user insists on browsing on their Windows 95 dial up 640x480 PC with an old version of Netscape Navigator? What if some user is trying to browse your website on their old Mono Nokia 3410 96x55 pixel Java phone with Opera Mini?

Browsers with Javascript disabled used to be the case in some corporate and public sector areas, although these numbers are getting less and less. If they are at work, then there's probably reasons why their Employer disabled Javascript and one of those reasons is that they don't want their employees online shopping whilst at work.

Another javascript-disabled area is older or less capable mobile phones. If they are using older or less capable mobile phones then there are a lot more issues than just Javascript e.g. screen size, bandwidth, connection speed, secure connections, testing etc and I for one definitely don't want to go down the route again of testing that complex ecommerce websites work on all the mobile phones out there, just on the off-chance I might get a visitor using one of these (my company used to develop Java games for phones).

By all means get your website working on the new breed of mobiles that deliver a desktop experience with Javascript enabled desktop-like browsers, like the iPhone, since these are increasing in popularity and usage.

If you do decide that having say an ecommerce website working and taking orders and payments with Javascript disabled is important for your client base, then check the whole process works with Javascript disabled. There's little point getting your ecommerce site working this way if the external payment page doesn't work with Javascript disabled. For instance, visit the PayPal Payments Standard external page with Javascript diabled and you get a message:

NOTE: Many features on the PayPal website require Javascript and cookies. You can enable both via your browser's preference settings.

In your particular case MovieTub, if your site is targeted at people into fishing, then perhaps the mobile browsing web user is more important to you, e.g. the person who is out fishing all day, who might check your website on their mobile phone. But I still see mobile phones in the near future as more read only web browsers, e.g. people might browse your products using a mobile phone, but wait until they get home before placing an order using a desktop PC, with all the input necessary for that order (address, card details etc).

For ecommerce in the near future, perhaps cookies is more of an issue with the EU cookie consent directive (depending on how "strictly necessary" is intrepreted and whether cookies are "strictly necessary" for ecommerce systems).

In the near future I reckon more ecommerce sites will make more use of Javascript for vital functions, especially as shop owners want to add vital functionality to their site or improve interfaces. For instance it is difficult (not impossible, but not nice) to get an interactive customise/design your t-shirt shop without some form of browser side scripting. Ajax sites are getting more popular too, even Google is doing it. Sooner or later people will get used to and see that Javascript autocomplete drop down list when they google search as vital functionality. Try using the adwords keyword select tool without Javascript enabled. Try using google analytics without Javascript enabled. Going the Javascript route nowadays doesn't necessarily mean that your Javascript links wont get indexed by Google either (but it needs very careful investigation).
 
Last edited:
Upvote 0
What you are are saying about order being important is true.

But in this particular case Google, and users, would disagree with you that the most relevant results are being returned.

If you compare the site's search results for:

asymmetric blades dual action

with google's results, you'll notice that google places different products top, the hedge trimmers which mention all of those words, the same types of products the site places top for the other search: dual action asymmetric blades.

Sorry I missed this reply.

This is to do with the weighting of searches as configured on this site. Different points are awarded for different hits... eg a single word match might get one point, but an exact phrase might get 5 points. Changing the word order will break an exact match and lose points there. The site also is configured to give good points for a "hit" in the product name (worth more than a hit in the text). So this again skews the results towards products that have some of the searched keywords in the title.

I don't think any ecommerce software will ever match Google's search ability - after all Google is all about search. I don't think any shopping cart software provider could concentrate as many resources on search as Google can (i know we can't!). Its also the case that Google has masses of search data to analyse. For example, in the above search it would help if the system is smart enough to know that asymmetric is an unusual word, and to increase the weighting for records that feature that word - that is what I'd do if I was building a search engine. But you can only really do that if you have a massive amount of search data to crunch (in which you can see its a rare word).
 
Upvote 0
I don't think any ecommerce software will ever match Google's search ability - after all Google is all about search. I don't think any shopping cart software provider could concentrate as many resources on search as Google can (i know we can't!). Its also the case that Google has masses of search data to analyse. For example, in the above search it would help if the system is smart enough to know that asymmetric is an unusual word, and to increase the weighting for records that feature that word - that is what I'd do if I was building a search engine. But you can only really do that if you have a massive amount of search data to crunch (in which you can see its a rare word).

Sure, no one is going to rent a cart for £20 a month and get Google style search performance built in!

But there is a tremendous value in a 'good as is reasonably possible' search function.

In fact, if our hosted solution search was improved to meet my ideal standard - I would pay extra. If it worked that well, a lot extra in fact!

As it is we're working on a fully integrated Google custom search. That also comes at a cost but it's a cost I'm happy to pay.
 
Upvote 0
This is to do with the weighting of searches as configured on this site. Different points are awarded for different hits... eg a single word match might get one point, but an exact phrase might get 5 points. Changing the word order will break an exact match and lose points there. The site also is configured to give good points for a "hit" in the product name (worth more than a hit in the text). So this again skews the results towards products that have some of the searched keywords in the title.

I don't think any ecommerce software will ever match Google's search ability - after all Google is all about search. I don't think any shopping cart software provider could concentrate as many resources on search as Google can (i know we can't!). Its also the case that Google has masses of search data to analyse. For example, in the above search it would help if the system is smart enough to know that asymmetric is an unusual word, and to increase the weighting for records that feature that word - that is what I'd do if I was building a search engine. But you can only really do that if you have a massive amount of search data to crunch (in which you can see its a rare word).
You don't have to program on-site search to be as good as google to get something that is more similar to google. There are quick wins to be had just by wrapping and adding value to the database's search feature and making better use of the database's full-text search query features and options. In the above case you shouldn't need to do anything fancy or complicated factoring in unusual words, it should simply be a case of configuring/programming the full text search query to give more rank to results which contain more occurrences of each of the different words, rather than more total occurrences of any of the words.

i.e. for the search term:

asymmetric blades dual action

if a page contains all of those words then it should rank higher than a page that only contains one of the words, e.g. blades, even if the blades page has more occurrences of the word blades in total (or has it in the title).
 
Last edited:
Upvote 0
Sure, no one is going to rent a cart for £20 a month and get Google style search performance built in!
The type of full text search I am describing above is present in:

1. easywebstore.net's aspx sqlserver-based (I think) ecommerce system, at around £25 a month, to check this feature search for guitar fingerboard rosewood on their client site amusic.co.uk

2. for php/mysql pay per month ecommerce services, this feature is also present on pay monthly oscommerce/zencart systems from the likes of paceretail.co.uk, rentacart.co.uk, etc

Done well, built-in search can also improve on google standard web search, by also showing product images as part of the results.
 
Upvote 0

Latest Articles