The Enstar blog

Creating a Private Cloud

Fri, 3 Jun 2011 18:57:00 +0000

Consumer websites and non-critical applications have been taking advantage of the benefits of cloud hosting, but medium-sized and small businesses and organizations have been reluctant, or unable, to follow this trend because of some severe downsides in Cloud hosting web applications.

First, though the up-sides: Low cost of entry, Browser based - access from anywhere, Work or Remote is same experience, Speed of Deployment, Automatic update deployment, Minimized infrastructure and Reduced costs.
Plus, some cloud providers offer value-added packages which let you manage your CRM, basic accounting, run blogs or create websites with shopping carts.

But there are some very nasty downsides:

Security - applications and websites hosted in a public cloud are grossly insecure. It simply impossible to know who is accessing your data from 'behind the scenes'. Just as you would not leave your bank account open for all to see, no more can you have transactional data, personnel records, accounting or other commercially valuable information stored in the public cloud. There is no data privacy. Your 'confidential data' could be stored anywhere in the world and viewed by one of thousands of technicians anywhere from India to Iceland to Idaho.

Continuity & Portability - who is to say that the cloud based service your business might rely on is still going to be trading next year?
What happens if they get taken over or just cease trading? Who is in control and who gets priority? What happens if you outgrow your provider? How do you migrate from one proprietary system to another?

Backup and redundancy - who has control over backing up your data and applications and how and when is it restored?

Unbounded costs - applications hosted in the public cloud can use automatic scaling, but that means your costs can spiral upwards rapidly.

Legal vulnerability - to search your office or in-house server the authorities must get a search warrant. To get to the information you've stored on a third-party's web servers, they only need a subpoena. Type of search can also happen without your knowledge and the Cloud provider may be barred from even telling you. As the data could be anywhere in the world, it is subject to seizure by any local law. And the data is always subject to local data protection and regulation laws.

The clearString Solution
Creating your own Private Cloud allows you to mitigate many of the risks and challenges of cloud computing, while leaving you to enjoy its many benefits.

clearString enables anyone to create and manage their own Private Cloud without specialist technical knowledge. You get upside without the downside of Cloud website and application development.

Getting data into your site

Tue, 14 Sep 2010 15:48:00 +0000

There's little doubt that data makes the difference between a 'me-too' brochureware site, and an active and useful internet presence. However that data isn't always where you need it to be – on the site. What we need is an easy way to import, lookup, or synchronise data from other sources, and make it available to the table/query/view system that powers the data handling in clearString neatComponents.

Today I'm talking about the component that manages that process, to give you a feel for what's possible. The 'External Lookup Data Feed' component can be pointed at an external source – typically an XML file or an RSS or Atom feed – and will pull the data found there into one or more tables within the system.

A simple example would be where you are doing a one-off import, to migrate a blog from a legacy platform like Wordpress: point the component at the Wordpress blog's RSS feed, map the feed fields to the ones in the table, and trigger the lookup. All done in moments.

If it's not a one-off import, but a regular pulling in of data you need, simply add a schedule, and it will repeat the process every so often – you decide the schedule – and it will take care not to duplicate any records already imported, but will update any which need refreshing.

Sometimes the dataset is too big to import, or simply unavailable in complete form. In these cases you need to perform an on-demand lookup, and it will pull in the particular records you need. For example, on a car parts website, you might want to let the user enter their car registration plate, and the lookup asks the national database which make and model it is – and thus present a catalog of the parts suitable for that particular vehicle.

Again, the lookup component can handle this, passing forward the search terms entered by the user in the request to the external database, having first checked to make sure the details aren't already known (to save time, and, if charged per request, money).

Sometimes things go wrong with external databases, so we also gracefully handle request timeouts, allowing for failover to backup feeds, and of course some lookups need to be to secure sites or require authentication to access them.

This has only given a quick sample of what's possible; for more information see: www.neatcomponents.com/external-lookup

Keep on rolling out

Fri, 3 Sep 2010 21:43:00 +0000

As we wind down for the long Labor Day weekend here in Connecticut, its a time to look back over this last week with a great deal of satisfaction.

The technical roll out of clearString's latest edition with the new 20351 version of the neatComponents engine went very well. Last month's developments in external data lookups massively increase the power that clearString brings to Internet developers. We're really excited!

Next week will bring a new set of challenges, I am sure - but I can hear lights being switched off behind me, and we have Hurricane Earl who is taking aim at us. So it's time to go.

Have a great Labor Day week-end!

Hello clearString

Tue, 31 Aug 2010 17:08:00 +0000

It's always great when a plan comes together, and all the more so when the plan has been several years in the hatching. Today's launch of clearString marks the end of the development and beta testing phase, and the start of a rollout to designers and business managers.

To recap, there's lots of ways of building websites, but once you get beyond a basic brochureware site, you're into the expensive territory of coding and programmers. Mention database, and most designers start screaming, either to themselves (watch their eyes), or out loud.

clearString changes all that - it's software package that makes it possible for people to make real Internet based businesses - without programming and without being trapped in predetermined template systems - by using new 'do-it-yourself' Internet applications.

If you think programmers are too expensive and you want to get more value from your graphic designers, look at clearString for a ten-fold efficiency gain (and no, that's not a typo, it really does make that much of a difference). So projects that normally cost $20,000 can now come in nearer $2,000, which means you can afford to do more, sooner. Take a look at clearString today, and be prepared to be pleasantly surprised.

New Internet presence for Mailtraq

Fri, 27 Aug 2010 16:27:00 +0000

In addition to clearString and neatComponents, Enstar provide distribution and support services for Mailtraq, the affordable alternative to Microsoft Exchange Server for small to medium sized organizations and I am pleased to see that the new website for Mailtraq is going live at http://www.mailtraq.com http://bit.ly/dDuhh1

It is a big improvement on the previous web presence giving more visibility to the most desirable features in Mailtraq, and splitting out the background feature description and general Knowledgebase onto its own site at http://info.mailtraq.com http://bit.ly/9XdJfK

A great job done by everyone involved. Congratulations!

New support system for clearString

Fri, 27 Aug 2010 14:09:00 +0000

It's great to see the new support system for clearString ready to come on stream. One of our recent targets has been to improve the user experience of our support services. The new system integrates license management, upgrade protection and support access into one handy on-line account so that all the information both the user and our support staff need to know is in one place.

Enstar was a early adopter of dispersed support, with our support staff working from their own homes. VoiP telephony, email and a heavy investment in Internet based technichal doumantation open for all have allowed us to keep standards high whilst keeping costs - and so prices - in check.

ClearString feature releases

Wed, 25 Aug 2010 13:37:00 +0000

One of the great things about clearString is that we don’t save up lots of features for months and then have a major version release, where everyone has to learn tons of new skills all at once. Instead, we release them as soon as they are ready.

However there’s always interest in what we’re currently working on, so I’ll be posting about that – so you can be thinking about how you can use new things as soon as they arrive.

How to win a website development contract

Tue, 2 Mar 2010 14:26:00 +0000

So you’ve got a lead to see a potential new client. They want a website, and they’ve heard you’re new and have some great ideas. So what’s your plan?

Here’s your cheat-sheet to winning their business:

Before you meet the client…

Understand your customer

- research them
- research their market sector
- research their local area
Be prepared to spend an hour or two on this. Even if you don’t win this job, you’ll be able to reuse the knowledge by approaching their competitors. Talking of which…

Understand their competitors

- Analyse their sites
- See how they promote themselves
After all, their competitors have been living that sector for much longer than the couple of hours you have. Figure out the tricks they’ve been using.

Then, on to the meeting itself.

Where to meet?

Try to meet at your client’s offices, not your own. However great your offices are, they’ll never meet their expectations (too small, and you lose credibility; yet too large and they’ll think you’re overcharging to pay for them – you can’t win that one). Plus, meeting at you clients allows you to learn more about them, and will prompt you to cover issues neither of you would otherwise have thought relevant.

Online demos

This is a harder one to call, but I’d try to avoid looking at websites in the first meeting, unless the client is really keen to show you things. The issue is that you can rapidly get drawn into the detail, when you would be better off learning about general issues and building a relationship with them.

Don’t flash your iPhone /iPad etc

Sure, it’s nice to have the latest technology with you, but that’s not what the meeting is about – you’re selling a website, not a phone. What’s more, you don’t want to get drawn into one of those philosophical debates about whether you should be a PC or a Mac. And if you have your phone with you, turn it off! There’s nothing more off-putting and rude than brushing the potential client aside to answer the phone.

Take notes

On paper. Remember, you left the iPad at home. You can type up the scrawled notes later, but in the meeting itself you can’t beat the immediacy and speed of pen on paper.

Keep focussed

You may be the first technically-capable person they’ve seen in weeks, so may want you to fix their printer, sort out the wifi and advise on why their computer is running slowly. Unless that is a (chargeable) aspect to your business, politely decline the opportunity to waste your time on it. However easy you think it is to solve, there’s always a catch, and that will be your fault.

Listen to your customer…

What is their motivation?...
- Find out what problems they are facing
- Ask why they feel they want a website at all

Give brief comments to show you’re engaged, but hold back from showing all your ideas at this stage. After all, you’ll want to think them through first.

Ask lots of questions - even if they aren’t related directly to the website. You need to understand their business to know what will work well on the web:

Get the rough financials:

- Turnover (A larger site means its worthwhile spending more on the detail and functionality)
- Gross profit margin (if the margin is large then there are opportunities for creative discounting promotions)
- Number of customers (if they are caring for just a few customers the emphasis may be different to handling thousands)

Find out how much they want to spend on the website

Clients may start out thinking choosing a web designer is simply a tendering exercise, where the cheapest bid gets the job, but you need to let them understand that price (or rather, value) is just one aspect – and that making sure they get the right website is even more important. So ask directly how much they want to spend on the website. Liken it to buying a car: where the dealership has different ranges to suit different budgets, and its pointless wasting time discussing high-end models if they can only afford something basic.

Understand their skills:

- Are they born writers? (If so, you could get them blogging etc, but if they aren’t confident with words, perhaps you could offer to do that for them)
- Can they take photos – good ones? (likewise, if not, you could offer that as an extra service)
- How open to new ideas are they?

Give them some homework:

They’re probably talking to a few other people before deciding who to give the job to. So you need to get them engaged with you as early as possible. Ideally now. So give them some homework. Ask them to come up with a list of five sites that they like, and five that they dislike. They don’t need to be in their sector, but they should write line or two explain their reason for including each one in the list. This will give you some concrete ideas about their design preferences, and allows them to do this in their own time, without wasting yours as they trawl the Internet (which is quite big these days!). Give them a deadline of a few days, so you can incorporate the feedback from this into your proposal.

At the end of the meeting

As the meeting ends, explain you may have some follow-up questions over the next few days as you prepare the proposal. (And make sure you come up with some – they show engagement).

Ask if they have any questions about you and your service, and explain you will write to them with a fully-costed proposal in ten days or so. Try to avoid predicting what will be in it: it’s all too easy to get cornered into giving a firm costing at this stage, and you must give yourself time to consider that. But you should be able to give them some comfort that your proposal will be in the general price-range they detailed earlier in the meeting.

As you leave the meeting

Hopefully it went well, and you’ll get the job. But, do you really want it? Some clients are more trouble than they are worth, and now’s the time to bail if you feel they won’t fit nicely with your portfolio. However, assuming you got on with them, you need to start thinking about the proposal document.

Next time – how to write that proposal.

Creating a social marketing strategy

Thu, 4 Feb 2010 09:30:00 +0000

No matter how good your website is, if no-one know it’s there, it’s a complete waste of time. So it’s important you have a strategy in place to spread the word. Naturally you’ll be doing all the standard cross-referencing: including it on your business cards, packaging etc, but if you really want to reach out to new people you’ll eventually have to think about online marketing.

Unfortunately this can be a minefield. The landscape is evolving, with lots of players like Twitter, LinkedIn, Facebook, Myspace etc all vying for attention, and each varying in their relevance depending on the nature of your business. Where to start?

There are two things I’d suggest for ‘day one’: See what your competitors are doing: if they’re already doing this, you can at least match that, and then see what else you can do. And, regardless of them, I’d start with some mainstays of online marketing: write a blog, post links to it to Twitter, and build a Facebook fan page.

Don’t rush at it though: despite the apparently trivial nature of some of these platforms, your reputation hinges on every public posting you make. Choose your language carefully, don’t act on a whim. Think about how things will read in the cold light of day.

Plan for the long-term. Writing a blog requires commitment: far too many blogs are started, and abandoned forlornly a few days or weeks later.

neatComponents provides some key technologies to help you manage and integrate all this: you can create your blog within your site: no need for it to be on a separate blog site where blog traffic won’t benefit your business. And you can get the site to store up and publish your blog postings on a schedule. Create a batch of postings in one session, and leave it to manage itself for a week or two.

It also integrates automatically with Twitter, giving you a smooth way to drive traffic to your blog – and to your site. Again, this can be done to a pre-arranged schedule: so a Tweet can be sent out when your blog posting goes live, and a reminder Tweet can be sent out a few hours later, catching a separate audience.

There are lots of people out there who will guide you through this whole process, but it’s hard to assess the professionals from the hopeless. Once business which falls squarely in the ‘professional’ category is Consoltis. They’re based in the UK, but work with clients worldwide, and put a lot of thought into providing a personalised service. They get a mention here as they are one of just a handful of “neatComponents Gold Certified Partners”, and they really know their stuff.

Your web, post-Amazon

Wed, 3 Feb 2010 14:00:00 +0000

Last time I looked ahead to a world where global players like Amazon overstep their role in the market, and a forced to take a back-seat.

Today I’m looking at what will fill this vacuum, and how you can get position yourself now to be able to take your fair share when it happens. In essence this boils down to two key concepts: ‘own your data’ and ‘structure your data’.

Owning your data means keeping control over it. Don’t simply post it to content aggregators like Facebook or Twitter, where it is easy to add the information, but hard or impossible to move it to another platform. (But it’s okay to use them to point people to your content).

Keeping control means you should own the website where the content is delivered: that way you will not be beholden to others who might decide to switch you off. Luckily self-hosting is easy to do these days, as the cost of hardware reduces you can lease a dedicated server for $100 a month, and there are even cheaper methods if you host from your own office.

“Structuring your data” is all about making is usable, both now, and in the future. If you structure it properly – using forms and a database to store the information, splitting it down into different fields, then you will be able to present the information in lots of different ways, including ways that haven’t even been thought of yet. But if you just throw the pages together at the beginning, they may look great now, but later on you will have a terrible time restructuring it.

So, if you’ve prepared for the post-Amazon era by taking ownership of your data, and structuring it sensibly, what can you expect? More players will enter the market to search through the structure of your data to make it available and usable in different forms. So if you are selling products, you might find your items listed in lots of places, with innovative ways emerging for the user-generated content (comments, ratings, companion-product preferences etc) to exist outside of the walled-garden of a single vendor. Naturally we’re unlikely to have heard the last of the present giants: I’d imagine both Google and Amazon will evolve into this aggregating role, along with players who currently seem to be in a totally different place: for example hardware companies like HTC, and hybrid hardware/content companies like Sony.

A valedictory for Amazon

Tue, 2 Feb 2010 11:00:00 +0000

There’s an interesting little argument going on between Amazon and the publisher Macmillan, which led to Amazon pulling all Macmillan title from their site for a while. Whilst I won’t repeat the detail of this particular case I think there are some important lessons for us all about the power of the global players on the Internet.

What we’re seeing here are two major corporates having a fight over how much money each will get from their ebook sales. In all this thrashing about, two sets of people are being forgotten: the authors, and the readers.

So what’s likely to happen next? Amazon will continue to try to become a monopoly in the marketplace, and become a monopsony in their purchasing, and thus grab the profits from more stages is the supply chain.

But inevitably the power this brings also brings danger: if Amazon can – at a whim – decide not to sell a publisher’s products, and if Amazon are the main purchaser, then can effectively censor what is published. Of course one would like to believe this would never happen, but we’ve already seen it done with Macmillan over a petty pricing disagreement. Imagine what would happen if Amazon chose not to list books that disagreed with a particular religious, or ethical opinion.

But despite this I’m optimistic. Not because I think the global corporates will discover a moral compass, but because the Internet is bigger than them all. As soon as they loose sight of the people that matter – the suppliers (authors, in this case) and the consumers, they will cease to be the best route to market.

So what will the post-Amazon “new world order” look like? I think it will be a lot more interesting – with individual suppliers and consumers having their own websites, and a ‘semantic web’ tying it together to prevent this becoming an unmanageable tangle.

The emphasis will be on ownership and control, with people distrustful of letting corporates run their sites, or even to host them. Global players – like Google – will have a role in tying this together, but just as the ISPs have had to accept they are simple pipes, they too will have to accept that they have very simple specific roles to play.

This won’t happen overnight, and there’s going to be a lot of pain in the meantime as the global players struggle to accept their new (reduced) roles.

For the rest of us, a race is just starting, to understand the value we can offer over the Internet, and who it is of value to. Just don’t expect that the best route to market is via the likes of Amazon. It might seem like a cheap short-cut today – but that’s what they want you to think.

Next I’ll be looking in more detail at the post-Amazon era, and outlining some techniques to make the most of it. Exciting times.

Cloud computing Oops

Fri, 22 Jan 2010 13:00:00 +0000

Sometimes the next big thing is too good to be true. Finally it’s beginning to dawn on the big players in the cloud computing market that their business models are holed below the waterline.

For once the problems aren’t simply technical, but more fundamental. Hence this week it was Microsoft’s lawyer who got wheeled out to plead with lawmakers for a Cloud Computing Advancement Act to change the world to help make it work.

Of course he’s not going to get very far with that. The Internet – and the ‘cloud’ – is global, and even if legislators in the US accede to his wishes, it’s implausible that the rest of the world will too.

The irony is that whilst the Cloud has the benefit of allowing rapid upscaling – and so is great for supporting the next phenomenon, such as Facebook or Twitter, yet almost every other website is of a fairly predictable scale. Sure, they might double their traffic over the course of a year, but in hardware terms that’s trivial to handle. Certainly there’s no need to sacrifice the flexibility and control you have by rebuilding for the Cloud.

So what can we learn from all this? The cloud concept has some great attributes – remote configuration, and reduction of the technical knowledge that should be needed in configuration. Perhaps it’s possible to retain these whilst avoiding the pitfalls of privacy, reliability, loss of control, and architectural compromises that go along with it. Sounds attractive to me. See www.cloudofone.com and let me know if you agree.

Create your own Press Release approval system

Thu, 21 Jan 2010 10:00:00 +0000

Last time I looked at a simple approval system for a blog comment – basically adding a checkbox that only the admin can see.

Today I’ll look at a more complex scenario, more suitable for a Press Release. In this scenario we want the Press Release to be approved by three people: one checking grammar, another the technical content, and, once approved on those aspects, a final approval by the CEO.

For this we’ll use three checkboxes, named: Grammar, Technical, and CEO.

To underpin this, the site needs to have four usergroups (in addition to the built-in Visitor usergroup): These we’ll name: Editor, Grammar-Editor, Technical-Editor, and CEO

For each usergroup we create a Data Entry Form, which shows the text of the press release, and, for Grammar-Editor, Technical-Editor and CEO, shows their respective checkboxes.

When the Editor has submitted the Press Release, the system sends an email to all members of the Grammar-Editor and Technical-Editor usergroups. No need to bother the busy CEO at this stage.

When either the Grammar-Editor or the Technical-Editor save the record, the system checks to see if they have both approved it yet, and only if they have both approved it, an email is sent to the CEO asking for final approval.

Once the CEO has approved it, the Press Release can be published. This is achieved by simply adding the criteria to the Query embedded on the page, to see if the CEO checkbox is checked.

Note that it probably isn’t sensible to add criteria here to ensure that the Technical-Editor and Grammar-Editor have approved it. This should have happened anyway as part of the workflow before it reached the CEO, but even if it hadn’t you probably want to allow a side-flow so that if the CEO wants to approve something and get it released, then it happens, no matter what anyone else thinks!

For those of you familiar with the free neatComponents system, it goes without saying that all this configuration is done without writing a single line of code, without any Linux haikus, and without breaking a sweat. If you’re new to the system see www.neatcomponents.com

Creating a moderated comments system

Wed, 20 Jan 2010 14:00:00 +0000

Here’s a quick trick that will make life easy when you want to keep control over what is be added to your site. I’m going to skip over the basic Content Management approval controls, which are documented elsewhere, and look at the form-based techniques that give you even more flexibility over your approval processes.

Let’s start with a simple typical scenario – a blog with a comments section, where you want to allow anyone to post a comment, but you want to approve (ie moderate) the posting before it shows up on the site. Later on we’ll extend this example to more complex setups, but this will explain the fundamentals.

So, for our blog, the comment postings will be stored in a neatComponents Form – with the usual fields of Comment text, Date, Posting (that it refers to), and so on.

The trick is to add a checkbox field, called say ‘Publish’, which is unchecked by default.

Then, when the Query that displays the comments is embedded on the page, add a criteria to only show records where the Publish checkbox field is checked.

Naturally it’s important that (a) the writer of the comment doesn’t see the Publish checkbox – or they could approve their own postings(!), and (b) the site administrator does get to see the Publish checkbox, to approve them.

This is achieved by using two Data Entry Forms, both linked to the same underlying Comment Form. The first one is public facing, and doesn’t contain the Publish checkbox, whilst the other does show it, and has its Permissions set to only allow the site admin to access it.

For added sophistication it’s a trivial step to add an email to the Public Data Entry Form that sends the admin an email to alert when a comment needs moderating.

In my next posting I’ll be building on this simple example to look at scenarios where an item (probably not a blog comment, but, for example, a Press Release) needs to be approved by several people.

For those of you familiar with the free neatComponents system, it goes without saying that all this configuration is done without writing a single line of code, without any Linux haikus, and without breaking a sweat. If you’re new to the system see www.neatcomponents.com

The point of ecommerce

Thu, 10 Dec 2009 14:31:00 +0000

What's the point of ecommerce?

Before you think I’ve lost the plot completely, perhaps I’d better clarify the question. If you have an ecommerce site, what should you measure to judge its success? And if you have a web designer or SEO specialist to hand, what do they think is most important?

There are lots of easy (and wrong) answers: you can measure traffic, you can measure length of visitor stay, you can measure how often people come back to your site. But interesting as these, and countless others are, they are simply stepping stones to the one metric that really counts.

But you’re ahead of me already. It’s revenue – the more you sell the better the site, so, thankfully a really simple one to measure. Well, nearly, but not quite. Revenue itself is great, but only if its profitable. So what you actually need to be measuring is profit.

Unfortunately measuring the profit from an ecommerce operation isn’t quite so straightforward – particularly if parts of the operation are shared with non-ecommerce routes to market, and you then have to apportion their costs fairly.

However, even if you can’t calculate the figures exactly, keeping the focus on the concept of profit, rather than any of the other technical metrics, will stand you in good stead. That’s the point of ecommerce.

Bounce rates

Tue, 28 Jul 2009 07:49:00 +0000

One of the simplest, yet often misunderstood metrics in website analytics is the 'Bounce Rate'

The Bounce Rate is the percentage of visits to your site that consisted of just one pagehit. The visitor arrived at a page, and that was it. No further clicks.

The tricky part is deciding whether a high bounce rate is good or bad. And to determine that you need to think about why you want people to come to your site. If you have an ecommerce site, where you want people to buy things, and to buy they will need to click through at least one page, then a high bounce rate is a bad sign. (But if it's an ecommerce site you probably want to be measuring the transactions themselves too.)

However for many sites, that are purely informational - say brochureware sites describing an organisation, or listings sites - a high bounce rate may well be a good sign. Think of it from the visitor's perspective. They've done a Google search and a page from your site has been thrown up in the search results. They click on it, see the information they need, and... well, that's it, they're happy. A successful visit, yet one which is counted as a bounce.

And this will happen more often than you'd imagine. Many, many visits to websites are to find the organisation's contact details, so they can phone them or post something to them. And if you have that information prominently on the foot of each page, they won't need to go clicking around trying to find it.

So if you can't simply say that a low bounce rate is good, what's the point of it? Well, for any particular site, it's worth watching the trend over time, as rapid changes in it can point to issues which need more investigation. For example, you may find that you have unexpectedly been included in the search results for some new keywords, but the pages people land on are not as clear as they could be - so your bounce rate increases.

Ultimately, the bounce rate is just one of hundreds of metrics available in tools like Google Analytics: they all have their uses, but often need to be viewed together to get any meaningful understanding of your visitors.

What were they thinking!

Wed, 22 Apr 2009 09:30:00 +0000

One of the hardest aspects of designing a site is making sure it meets the needs of your audience. And regardless of how much research you do beforehand, there will always be surprises when the site goes live. People are strange. They act in unpredictable (and sometimes, dare one say, 'stupid' ways!). So you really need to know what they are thinking when they visit your site.

One good way of getting a view of this is to see what they are searching for. You can log all their searches on your site, and then look through to see what they thought your site wold contain. What you're really interested in is where the searches failed - that is, where there were no results shown to them, or where there were a massive number of results shown. Sometimes this will be because they can't type, and mis-keyed their search query, or it may be because they used different terminology than that on the site. Either way, you can rescue future such searches by adding in words on your pages to match. (For typos you can use the Keywords field, so your page isn't visibly contaminated by mis-spellings.)

In neatComponents you can ask for searches to be logged, and then use the Search Log Datafeed Component to retrieve the logs for display using a Query.

Not only are the search logs useful to improve the user experience on your site, they also provide a fascinating insight into your users, and will likely set you off developing the site in directions you had never before considered. Happy logging.

Avoiding repetition

Mon, 20 Apr 2009 09:00:00 +0000

I really hate having to do the same thing over and over again, and I know I'm not alone. All too often on sites you need to make a change to lots of different pages maybe changing the font or the background color, and you really don't want to have to do each one individually.

The solution to this is a property called 'inheritance'. In neatComponents all the styles, settings and permissions (ie everything set in a page's Behavior Editor) can be inherited. This means that if you set the value for one of these settings at the top level of the site tree, they will cascade down and be applied to all the child pages in the site.

Naturally, there are occasions when you want a page to be different, and setting the value on the behaviour Editor for a specific page will override any value it would have inherit from above.

So, when styling a site, or setting permissions for pages, always think about how many pages will be affected, and try to work as high up the site tree as you can. That way you'll avoid repetition. What's more should you need to change the site later on, you'll only need to make the change in that one place.

You can learn more about inheritance in neatComponents.

"I want to be a web designer"

Thu, 16 Apr 2009 08:12:00 +0000

I was recently involved in a panel discussion where a questioner asked what they should do to become a web designer, and what certification courses they should take.

The consensus on our panel was that certificates in this sector are worthless: employers don't trust them, and are more interested in seeing a real-world portfolio showing experience. Which is, of course, a bit of a catch-22 for someone getting started. We suggested they should start by finding local not-for-profit organizations and offering to assist with their websites, and build up a portfolio that way. This also has the benefit of providing a local contact-base which could lead on to some paid contracts.

The other point we made on the panel was the distinction between a web 'designer' and a web 'developer'. Web designers are primarily focussed on the graphics and the look of a site, whereas developers are more concerned with the underlying mechanisms: databases, user permissions, security, ecommerce, etc. Most people want to concentrate on the graphics, and are thwarted by the need to program to do the other aspects. Naturally this is where neatComponents comes to the rescue, as it provides all these technical elements without the need to learn any programming languages - and of course there is a free license designed to enable aspiring web designers to get started with the minimum outlay.

Being logged in, and out, at the same time

Wed, 8 Apr 2009 11:09:00 +0000

When you're building anything but the simplest site you're likely to be faced with the problem of seeing what the public view of the site is, whilst needing to be logged in to make changes. You could keep logging out and back in again, but that's time-consuming and tedious. So here's three ways to be both logged in and out, at the same time:

Use two computers!
Use two browsers on the same computer so login with Internet Explorer and be logged out with Firefox. (Login status is maintained by cookies, and they aren't shared between browsers)
Use two domains set the site to listen on two different domain names, and login to one of them. (The login status cookies work on a per-domain basis)

Whichever method you use, having multiple monitors to work with makes life a lot easier too!

Helping you sleep at night

Fri, 3 Apr 2009 11:09:00 +0000

Here's a thought that might save your life. Well, your career. It's all to do with that mundane task that's never quite important enough to do right now. Yes, it's the dreaded backup routine. Though today I'm concentrating not just on backups, but the equally important restore process.

So first off, ensure that your hardware has some redundancy. A RAID disk array is a must-have. Disks die all the time, and without notice. Make sure you have an alert setup to tell you when a disk has died, otherwise after the first disk has died you can be running under a false sense of security on the one remaining disk.

With the RAID in place for minute-by-minute protection, we now need to step back and protect against the issues that could require you to change hosting provider. These could range from a total machine failure, to the colo losing all their connectivity, or even going out of business. In these situations its important to already have an offsite backup of the site: by definition, you won't be able to go and take a backup when things go wrong.

So, how often do you backup your website? Monthly? Weekly? In the 'old days' when websites were pretty static brochureware, you could get away with a monthly backup, and after a restore it would be little hassle to redo any changes that might have been applied in the meantime. However with data driven websites you can't backup often enough data is changing all the time - and with user-generated content you may not even be aware of it.

There are two ways to get a backup. You can Export your site, via the Layout Manager, or if you have multiple sites you can use a backup script to grab them all at once.

Once you have your backups, it's time to rehearse what you would do with them in an emergency. Prepare a machine (it can be a Virtual Machine) with a fully patched operating system and the latest version of neatComponents, and restore your backup into it.

The other thing to be aware of is that unless you can quickly change your DNS records to point traffic to the new server, your site will remain invisible to the world. So, make sure that your domains' DNS records are hosted somewhere different from your neatComponents server, and make sure the TTL is set to something low, say an hour. That way, after changing the DNS to point to the new server everyone will be looking at it within the hour.

Feeling Agile

Mon, 23 Mar 2009 11:08:00 +0000

If you've ever studied computing at university you'll have come across some tedious methodologies like the 'Waterfall Model' or the PRINCE2 project management standard. One of the more recent entrants into the field is 'AGILE' development, and as the name suggests, it is all about getting results quickly, keeping the customer in the loop, and providing a working system as early in the development process as possible.

AGILE development is a team-oriented approach, with several developers working on a site at the same time, watching what each other is doing, and fixing issues as they arise. neatComponents provides an ideal platform for this type of working, as it is inherently multi-user for the site designers, with full page and record-locking, and it is quite happy for those developers to be geographically dispersed. The rich permissions system allows the client to be shown parts of the solution as it is created, without the outside world seeing it too.

"Look at my site!"

Thu, 19 Mar 2009 11:08:00 +0000

I'm often asked to cast a critical eye over websites, and suggest how they can be improved, or 'taken to the next level', as marketing people like to say. Aside from the usual mantra of making sure the content is up to date and the typos have all been found, there's one technical trick that can give a website that extra quality feel.

The trick is 'AJAX'. The term means different things to different people, but in essence it's all about making changes happen on a page when the user does something, without loading a completely new page. For example, if you have a list of popular products on your homepage, you can use AJAX to display a box about each product on that page when the user rolls their mouse over each product. It works for the user as they don't loose track of where they were, and it happens far faster than loading a completely different page.

You can learn more about neatComponents support for AJAX here.

Measuring site traffic

Tue, 10 Mar 2009 11:07:00 +0000

Everyone wants to know how busy their site is, and there are plenty of ways of finding out. The trouble is, they all seem to give different answers so who is right?

For example, neatComponents provides a component to provide usage statistics the "Performance" component. This shows, for the whole site, or by section or page, how many page hits are received each day. You can also integrate with Google Analytics, and use their reports to see the same sort of information. However whilst the trends from one time period to another will agree, the precise numbers almost certainly won't. Yet both neatComponents and Google Analytics are faithfully reporting what they see. What on earth is going on?

neatComponents will count those pages which it delivers, but it won't include those pages requests which never reach it ones which are provided by intermediate caches at ISPs, or from the browser's own cache. All those will be included by Google Analytics.

Google Analytics, however, will only see those page hits from browsers with Javascript turned on, and which allow Google Analytics to analyse their hits. So browsers on mobile devices are likely to be particularly under-counted by them, but included by neatComponents.

So no-one gives the complete picture, just the part they see. However in most cases it doesn't matter too much concentrate on the trends and so long as they are upwards you're doing fine.

A quick tip. If you haven't been measuring your site traffic, but want to start now, simply add the Performance component to your site. You'll find that the site has actually been measuring the traffic all along, so you'll be able to see all your historical data too. And that's something you can't do when you turn on Google Analytics!

Canonical

Fri, 20 Feb 2009 11:07:00 +0000

It's not often that Google, Yahoo and Microsoft agree on something, so when they jointly launch a new standard it's worth taking a look at. In this case it's a really simple standard just a tag but it's going to make a big difference to search engine optimisation. What's more, this isn't some pre-announcement of some future development, it's here now, and neatComponents already supports it.

So what is it? The canonical tag is a simple way of saying that two pages are actually the same. For example, say you have a page www.example.com/27 and you have a DirectURL on it of www.example.com/news. In the eyes of the search engines these are two different pages, and any PageRank will be split between them. Worse, they may spot that the content looks the same, and penalise you under the misapprehension you are trying to trick them with 'duplicate content'.

The Canonical tag come sto the rescue by allowing you to say that both URLs are really the same page. The canonical tag allows you to specify the URL you'd prefer them to use (so often you'd make it the DirectURL). With the Canonical tag in place, the search engines from Google, Yahoo and Microsoft will understand that both URL are the same page, and aggregate all their information about them inward links, PageRank etc.

You can learn more about neatComponents support for the Canonical tag here.

Learning the system tricks of the trade

Mon, 5 Jan 2009 11:06:00 +0000

Getting to know any system, even one as easy as neatComponents, has its tricks, and whilst you'll find them documented in the 'official' places, I'll try to draw some to your attention here.

Many of these are obscure settings, or way you use them together in combination you might not have thought of, but today I'm looking at the bigger picture. If you're trying to create your first site in neatComponents it can be a bit of a learning curve. So, concentrate on the learning for a bit. Leave working on your site for a bit, and find someone else's site out on the Internet, and try to replicate it as closely as possible using neatComponents.

This exercise is really useful, because it forces you to find out exactly how to do something, rather than working around your initial lack of knowledge of the system. As you create your copy, you don't need to worry about trying to visualise how the site should turn out, as you have a working original in front of you. Any time you wonder how something should look, or behave, simply turn to the existing site. Like that. And whilst you doing this you don't have to worry about creating graphics for the site, as you can copy them from the site you are working from.

When you're done with the site copying exercise, return to your own site. You'll find it much easier, as you can apply all the techniques you've just learnt.

Finding your nearest hosting provider

Tue, 2 Dec 2008 11:03:00 +0000

I'm often asked by prospective neatComponents users where they should host their servers. I could easily reel off the names of some decent co-location specialists, but that's not the best way to approach the problem first, you need to know what you need from your hosting. And armed with those answers, you have a chance to sensibly assess the options.

I'll be coming back to this whole issue from various angles in future, but today let's start with bandwidth. Most colos offer unlimited bandwidth, or limits so high they are effectively not there. But bear with me, there's a point to this question. First off, we need to know how many page hits you get per day, and then we need to know how big in terms of bytes those pages are. We don't need to be totally accurate with these figures, as we're only after a ballpark answer to demonstrate the order of magnitude.

So imagine you have a site which gets 10,000 page hits a day, and with 100KB of content per page. That works out at 1GB of bandwidth per day, and is typical of an established website for a successful small business.

Now back to the question: where do you put your site? You can see from the numbers above that almost any colo will be able to provide suitable hosting 1GB of bandwidth is 'nothing' for them, but to get started, there might be an even cheaper alternative: host it yourself! As long as your ISP provides you with a static IP address (ie one that says the same each time you reconnect) you should be able to host from your own office. Take a spare machine (or create a Virtual Server on one), install Windows 2003 Server and neatComponents on it there are free trials of both available, and you have a start-up server for no cost at all.

You still need to think about backups, and obviously if your internet connection is unreliable it's not going to work well, but this is a great way to get started without any financial outlay.

Once you outgrow this, you can easily export your neatComponents sites and move them to a 'proper' collocated server.