The Beetrieve Buzz

They grow up so fast.

Beetrieve, in its pre-3.0 life, was a great app, functionality-wise. It was just not sleek. Like a teenager, it was eager and had great ideas, but looked a little... unproportional.

All that's changed now.

Enter Beetrieve 3.0. We've completely ripped the old look to pieces, gone over every pixel, placement and alignment and rewritten its design from the ground up. The fundamental layout is still the same, because it always worked great, but instead of a fixed size window, awkwardly boxed inside your browser window, the browser window now is the app window. Every element floats to position itself effortlessly within this window, whatever size it is (click on the images to embiggen).

This is more than an improvement in looks: The new layout yields more room to the participants list, the invitation email body text box, and so on. It is in fact a really big deal, and it means you can decide how much you want to see of the admin interface. This new layout even works better in sizes smaller than the previous boxed window.

Furthermore, we've gone over every interface element and added definition, clarity and just plain niceties in countless places: All admin panels, the survey list, the web shop, the users list in Beetrieve Enterprise, the Account panel, surveyCanvas, etc, etc.

But, if you think Beetrieve 3.0 is all about layout and pixels, you'd be mistaken. The other big news in Beetrieve 3.0 is the new Participants Panel.

When you work with email participants you'd like as much info about them as possible, and you'd like to be able to work on them, like resend an invitation, disable any that don't want to participate, and so on. You could get and do quite a bit of that in the Participants Panel before, but in Beetrieve 3.0 we have just cranked it up so much more.

This is from a real life Beetrieve 3.0 survey. After survey start, each participant gets a small status square in front of their email address. You can see what the half-green mean from the illustration (you see these tooltips when you hover over), and from that you can infer what the solid green mean: The participant has filled out the whole questionnaire.

White with green border: Email has been delivered, but participant has not opened the questionnaire yet. Red? You guessed it: Email delivery failed. Grey means the email is currently in queue for sending. Orange/yellow? You've disabled the participant.

All the squares are clickable. Clicking on a square with any green disables the participant. Clicking on a red one tries to send the email again (you can try sending to all failed at once with a command in the menu). And of course, clicking on a yellow square re-enables that participant. Pretty neat!

But that's (of course) not all. You've already spotted the 'More…' buttons next to each square. Here's what that does, for a failed email:

You get the email address, editable, and the exact error message that the email system generated. Now you can easily determine why the message failed and correct the situation, if possible. Just hit 'Try again' when done, and the email will be sent again.

Also, and that is there for all participants, you get that participant's link to the questionnaire (unless the survey has been started as anonymous, in which case the status squares also just show either sent or error). The link is a Beetrieve Enterprise only feature.

The improvements in the Participants Panel adds a whole new level of power for those that take their surveys seriously.

And seriously, with Beetrieve, we want to make creating surveys fun for people who take surveys seriously.

So what happens when someone opens and fills in the questionnaire when you have the Participants Panel open?

That's the other other big news in Beetrieve 3.0: We've removed every 'Update Status' button!

The whole web app is now just simply live, and updates automatically. Such an obvious thing really. It works great, and there is something incredibly satisfying about just sitting there watching as those little squares turn from white, green-bordered, to half green, to solid green, one by one, as the answers tick in. Of course, if you are more into numbers than colors, you can always switch to the Results Panel and watch the number of started and completed answers counting up in the same way.

Other new features? A few - the one to mention here would be the new and very handy survey info box accessible by clicking the cleverly named 'Show info' button above the survey list. Use it to add your own notes to any survey, for reference. These will then also show in the Survey Report. You can change a survey's name in the info box, as well.

See more screenshots, a few demo videos and more at Beetrieve.com.

So, can Beetrieve make it fun to create and run surveys? You tell us. Sign up and give it a try, then sound off in the comments, to customer (at) sonorait.com or use our contact form.

Over the weekend we released the newest version of our survey webapp, Beetrieve 2.7, to our servers.

The focus of this new version is to deliver a better email subsystem and subsequently to make sending test-, invitation- and reminder emails a better experience for all our customers.

What you will see, when using Beetrieve with email address participants, is that sending out invitations and reminders will happen faster and more reliable, that the feedback in the Participants Panel is consistent, and that updating the status in the Participants, Startup and Results Panel is much faster.

Also, now you will get feedback from starting a reminder email batch, in the form of seeing the emails being queued, and updates in the Paticipants panel.

We are already hard a work on the next version, which will be major. You have much to look forward to!

It started this summer. Every indication was that the webapp in itself was able to stand its ground against the best. But there was something lacking.

So we took a long hard look at the brand we had and saw that it wasn't up to snuff. We communicated in a way that was too technology-focused, telling our visiters lots of details and not focusing on the core of why we had made the webapp in the first place: putting a professional, enterprise solution in the hands of everyone that could use a browser.

This is a trap that countless others have fallen into before us: the product was developed by us, we are tech people with a vision about user friendlieness, that the product ends up having, but that we don't know how to communicate.

The first thing that had to go was the previous name. The name we landed on, Beetrieve, is short and sweet, and has obvious connotations. Along with the name there is the new mascot, who for the time being is unnamed. The website was next, and a lot of work was poured into redesigning it.

We are very proud of the result, and we hope you like it!

As you can see, there have been som changes here lately, and this blog has changed its name to 'The Beetrieve Buzz'.

Why? We have been working really hard this autumn to rebrand and relaunch our survey webapp, up until now called Laguna Survey. I am sure you can infer where this is going, but it is a little early to announce anything official just yet.

Watch this space for some exciting news later this month!

We are very happy to announce a brand new version of Laguna Survey, now live on our servers!

For this version the number one thing we have done is to completely overhaul the Result Panel and add a new great new feature: A results report that collects the results from your survey into an easy to read, comprehensive report.

The Results report runs from the first to the last question and displays multiple choice question results in tables and text from textual answers. It is great for getting a readable summary of the results for reference and discussion.

The report contains a few tools at the top; options to display the text you have entered into the questionnaire in surveyCanvas, and to number the questions in the report. And, in Laguna Survey Enterprise, a tool for including only a subset of the results in the report, based upon the participants' answers. See the manuals for more information.

Other enhancements in Laguna Survey 2.6:

You can now serve the questionnaire to your audience embedded in another webpage. When creating a new survey, the 'Show link' command will display the necessary html code for embedding a survey. See the manuals or help section in Laguna Survey for more details.

We have now introduced the ability to send HTML formatted email invitaions and reminders from Laguna Survey. The mechanism is simple but effective: You activate it with a single checkbox in the Startup panel: Check it and Laguna Survey will use the header you have created in surveyCanvas for the questionnaire, at the top of the emails as well. The text you enter for the email message in the Startup panel will be displayed in the HTML message as you expect, and will also be sent as plain text for those email clients which does not understand HTML. The message itself will display in an 800 pixles wide box (with the header image scaled down to match, if necessary), much like how the questionnaire itself is displayed.

We find that this works best if you create an 800 pixels wide image with the height of 60-90 pixels, but any image will do, and the background color will be included, have you specified one in surveyCanvas. The setting is ignored if you haven't made a header in surveyCanvas.

Supersecret, undocumented feature: You can now employ some simple formatting to the email message: Use codes to get [b]bold[/b], [i]italics[/i] and

in your message. No GUI is exposed for this in this version, so you'll have to enter the codes manually. Aren't you glad you read this?

We have switched on SSL as the default for participants for this version, which means the questionnaire is always served over a 256 bits encrypted line, just like Laguna Survey (the webapp) has been for some time. Note that the link that Laguna Survey produces still has the http:// prefix, the switch to https:// is done on first access to our servers. We are still investigating the optimal configuration for this.

The chart showing answers per day in the Survey report now shows both started and finished answers. More ready-made info for you.

The list of surveys now shows the newest at the top. This seemed like a better arrangement, especially if you have many surveys and the list starts to scroll.

Oh yes, accompanying this realease is a design and contents update for both the webapp and our public pages.

We hope you enjoy the extra spit and polish!

Running a survey is not as straighforward as you might think when you set out to create your first. Laguna Survey might make the creation easy, but there is so much more to it.

Overall, your main concern when running a survey should be to get as much quality data out on the other end as possible. Both quantity and quality is important: If the quality is not there, it means you cannot back up the validity of your results. You might have asked the wrong questions, the questions have not been formulated well, you have not explained the survey properly to the participants, etc. If the quality is there, you want to ask the right people, and you want as many as possible of those people to answer.

There is too much to go through in a blog entry. This is simply meant to point out a very specific mistake that we have seen many people do.

This is the list of element types that you can insert in Laguna Survey's questionnaire module, surveyCanvas™ (the entries marked with (E) is Enterprise version only):

What we often see is that users choose to use the "Multiple choice - multiple answers" question type when the question actually calls for a "Multiple choice - one answer" type. The first of these two produces a question with check boxes, boxes where it is possible to simultaneously check more than one box/answer alternative. The latter yields radio boxes, where it is only possible to select one such box.

The first is approperiate to use when each participant may be correct in ticking off several of the alternatives. For instance, the question:

"What pets do you own?

Parrot • Dog • Cat • Platypus • Turtle • Ferret • Pig • Other"

is one where the participant needs to be able to check more than one box in case he/she owns more than one pet.

The second, with radio boxes, is the correct choice when the answer alternatives are mutually exclusive. For instance with the question:

"How old are you?

10-19 • 20-29 • 30-39 • Older than 40"

the correct question type is "Multiple choice - one answer" because only one alternative can be correct for each participant.

It is an unfortunate mistake to mix the two, because it completely undermines the credibility of the results from such a question.

Why do people mix them up? I think it is mostly a simple case of not having really understood the difference. Also, it is a challenge to communicate the precise operation of a particular question type in a short line in a menu like the one above. Do some people misunderstand the text to mean that the question can only produce one result over the entire survey? It is possible, I guess, especially if you are new to surveys and Laguna Survey. We have always designed the interface with the intention that it should not be necessary to read any manual to start using the web app. No doubt this can be a challenge in some details, like this.

If you have a suggestion for the text entry for the question types in this menu, that you think is clearer in communicating how they work, sound off in the comments!

Finally! Proudly presenting Laguna Survey version 2.5

We've been working on this new version for a long time, and it feels great to finally unveil it. The biggest news is the one that you'll see on our home page: We've split Laguna Survey into to different editions: Laguna Survey Standard and Laguna Survey Enterprise. Both editions build upon the great foundation that we had in Laguna Survey 2.2.

There are a lot of changes, so let's get right to it. These are relative to the previously available Laguna Survey 2.2

• Changed the price model to be 100 free answers for the account
• Laguna Survey Standard and Enterprise are now served over an 256 bit encrypted connection
• Added questionnaire templates, available from the Insert menu in surveyCanvas.
• Added a 'Jump to page' pull-down menu to the questionnaire preview
• Better validation for the Multiple choice - multiple answers type questions
• Close surveys so that hey don't accept new answers
• Import paticipant email addresses from a CSV file
• You can now set the button text on the last 'Next' button in the questionnaire navigation

• Everything new in Laguna Survey Standard, except pricing model (Enterprise is sold with subscriptions)
• More templates, for the enterprise segment
• Several new question types, including multiple choice questions which allows you to use images as answer alternatives
• The report now includes a graph showing the number of minutes the participants used to complete the questionnaire
• The window showing result diagrams now includes more tools for analysis (see the reference manuals for details)
• Send the participants to an URL of your choice upon survey completion

That are only the major functional updates. In addition, there are an abundance of minor updates and design tweaks. For instance, there is a new surveyCanvas Tools panel:

The buttons will make it much quicker to access editing functions for the questions/elements you put in your questionnaire. They have tooltips showing their functionality, but In case you're wondering, they are, from the left: Delete, undo, redo, move up, move down, move to end, duplicate, preview, print, and export.

Of course, as you will see, all of Sonora ITs web pages have a brand new design. We hope you like it, we've worked hard to design a modern-looking site approperiate for a web app like Laguna Survey.

Of course, we've updated all our documentation, as well. The Laguna Survey User's manual is better than ever, and our reference manuals are as always available from the Help menu inside Laguna Survey, and they are full of new illustrations from the updated applications.

Have fun with the new version, we hope you like it!

God sommer til alle våre kunder og brukere av Laguna Survey!

Det er en stund siden siste innlegg og siste oppdatering av Laguna Survey. Det er rett og slett fordi vi siden den gang har jobbet hardt med den nye versjonen, som blir en stor oppdatering på mange måter. Det tar derfor litt lengre tid enn tidligere før alt er klart.

Det er ikke nye mer vi kan kommunisere for øyeblikket, men vi gleder oss til å slippe en helt nye versjon av Laguna Survey, og jobber så fort vi kan for at den skal bli ferdig.

We just released Laguna Survey 2.2 on our servers.

This release updates functionality in different places in Laguna Survey, and seeks simply to update aspects of the program where there was room to improve. The updated features are:

• In surveyCanvas, a new element 'Video'. You can now (for starters) place YouTube videos directly in your survey, by cutting and pasting the embed code for a video from a YouTube page into surveyCanvas.

This should be great for marketing campaigns, product surveys etc.

This is a start. Even if it is quite easy to upload and use video from YouTube, we want to give our users more alternatives, both to the sources of video and the format. This will come in a later update.

• More powerful branching. Before, when setting up branching, you could specify a set of conditions and one target. Well, now you can specify multiple targets, in the Branching panel, for the question type Multiple Choice - one answer (radio buttons - the only question type this is feasible for).

This has been a request from some of our larger customers, and we responded. It makes it possible to set up very powerful rules for branching, and we frankly think this is one of the most powerful branching systems out there, and at the same time it is actually easy to use!

For this question type, you can specify as many targets as there are answer alternatives (if you have that many pages), including the 'Last page' and 'End Survey' alternatives.

• Sticky preferences. This has been a long time coming, we know, but now Laguna Survey at last has a flexible system for handling preferences that survive from session to session. The first thing we have implemented is remembering the states of the 'Autoscroll' and 'Show element frames' in surveyCanvas, and Autoscroll got it first for a good reason.

Autoscroll is a great feature when you start using surveyCanvas, because it always brings the currently selected element into view. So, it makes very good sense to enable Autoscroll by default.

However, some of our larger customers have started producing very advanced surveys, with the number of questions and pages reaching the 50-70 range. Now, Autoscroll starts to become a nuisance, because you have to wait for the survey to come down the line (an XML structure sent by way of AJAX) and then for the browser to render same, and then the autoscroll will kick in and scroll the page to the bottom of the survey, to the last question. This can take many seconds for large surveys, and is not optimal, especially if you know that you are actually going to change something in the beginning of the survey.

This situation is now remedied in a simple and effective way. Just uncheck Autoscroll, and then it is off. Until you choose to switch it on again. Next time you open surveyCanvas, no Autoscroll. So you can start working immediately.

• This version sneaks in a new element that is only visible to our subscribers: A Text line table. Although not a major addition, it is a very handy element, and makes for a more effective survey in specific situations. You can try this element out in the demo.

This is our way of starting to walk down a path that we think will lead to a standard version that will contain more or less what you see in Laguna Survey today, and a premium version that will contain more of the features that is important for a certain segment of the market. That is all we can communicate at the moment.

• We have rearranged and added commands in the Participants panel, and put them all into a command menu. We are very pleased with the result, check it out for yourself!

We've also done some major work to make Laguna Survey more robust.

There are certain characters that any webapp has a hard time swallowing as user input, especially for an advanced app like Laguna Survey, where user input at one time or another might be parsed into XML, sent in emails, added to mySQL tables, handed over to a Java servlet as labels to use for rendering a chart, output in Excel and Word files, and generally thown around by AJAX. At the same time, the app should be able to handle such 'unexpected' input without any problems. Before this release we went hunting for situations where problems had arisen for our users, and then tried to take a holistic approach to fixing it. We think the result in Laguna Survey 2.2 is that it is very robust indeed.

Some more specific bugs have also been dealt with:

• The Drop-down menu element didn't work properly in the browser Firefox.
• There was a problem resetting the Minimum/Maximum that can be selected in question type Multiple Choice - multiple answers.
• In some rare cases the diagram for Multiple Choice - multiple answers could fail to render correctly in Result Diagrams when branching had been used.

We are pleased to announce that we just released Laguna Survey version 2.1 to new and existing customers alike.

In this version we have focused on surveyCanvas and strengthening its feature set and interface. The headline features for this version are:

1. Undo/Redo
2. Insert/Edit menus are now true command menus
3. A redesigned Global options panel that looks and behaves great

1. That's right, surveyCanvas now sports a five-level undo/redo buffer, and it works just like you would expect in a desktop application.

In a way, this is a strange feature to tout. Any ol' application certainly has undo functionality, right? Yes, and it's something we have come to take for granted. In desktop applications, at least.

The situation is different for web applications. For one thing, applications running inside a browser do not have access to application menus. Another aspect is that often, the operations that you would need undo for is taken care of by the browser undo function. For instance when entering text in a email message in Gmail or other web based email software: If you type something that you shouldn't have, hit the browser undo and the text in that text box reverts to its previous state, courtesy of the browser's undo buffer, not the web app's. In other web app contexts, undo simply doesn't make sense for one reason or another.

As a result, it seems that undo has never really made it into the mainstream of web applications, regardless of whether it would be useful or not.

surveyCanvas, and any other web service for producing questionnaires, is a perfect example of a web app that really benefits from undo functionality. Here, the browser has no grasp of what is going on, conceptually speaking, as you build your questionnaire, so it cannot provide any sensible functionality for undo in such a context. So it is up to the web app.

But you have to build your questionnaire-building web app carefully for it to be able to provide undo. in SurveyCanvas, the questionnaire exists as a Javascript object on the client side and an xml structure server side. The XML structure server side is authoritative and is where the undo buffer is kept and synced back to the client side on undo.

So, while undo/redo is something that we'd like our customers to take for granted, we are still happy to have it nailed down in surveyCanvas in this version. Now you don't have to worry if you accidentally delete the wrong element, that one with the long list of answer alternatives of car dealerships in your area.

2. The insert and Edit menus behave differently in this version. Well, there wasn't any Edit menu prior to this version, only buttons that executed the various edit commands. We added several commands in this version, and what we found was that the Tools palette was becoming too cluttered again.

The obvious solution was to gather all the commands into a menu. Menus on the web are not really meant for executing commands like in an application menu, where the commands execute the instant you select something from the menu, but this is what we have done with both the Edit and the Insert menu in this version. We hope you like it! See the linked blog entry above for more info on this.

3. As for the Globals Options panel, we thought it was high time to give it a make-over. The new version takes up far less space in its default state, can be expanded and collapsed at will, and works the same or better, especially the colour wheel. By the way, we use Steven Wittens great Farbtastic jQuery library to provide the color wheel.

Other enhancements in this version:

• 'Duplicate' and 'Move to end' in the Edit menu in surveyCanvas. Together these make it easy to build new pages in your questionnaire based on existing ones.
• New option to toggle the Question Area autoscroll on/off in the Tools panel.
• Support for .csv files when importing participant data.
• Possibility for exporting your questionnaire as a Microsoft Word file. This feature is accessible in the window that opens when pressing "Print/Export" in the Tools panel of surveyCanvas.
• The Results and Result Graphs pages now show "wait" graphics while the results pages are being built instead of just showing a blank page.
• Better functionality to stop indexing bots opening a survey from an open link posted on the web and thereby counting as a participant.
• The Laguna Survey User's Manual has been updated to reflect the changes in version 2.1.

Bugs exterminated:

• Pressing an update button in survey admin would just grey it out if the current session was expired. Now the user is correctly logged off with an appropriate message.
• Fixed a situation where clicking any of the links in the 'Finish' page of a survey in Preview would open a Sonora IT webpage inside the Preview.
• Fixed 'no email provided' from showing up in the email input field in the Test panel resulting from using the open test link.
• A couple of localization errors were fixed.

Known issues:

• The colour wheel in surveyCanvas does not correctly update its view in some versions of the browser Opera. It does however update the colour correctly, and the technical functionality is not affected by this bug. We are investigating this issue.
• The Text box input in the Options panel does not scroll its contents correctly in the browser Opera in cases where there is a lot of text in the box. We are investigating this issue. The technical functionality is not affected by this bug.

Also, we've made some significant changes under the hood in this version, which will serve us well in taking Laguna Survey forward in the months and years to come. Stay tuned!

As we work on updating Laguna Survey and adding new features, there is one question we constantly ask ourselves: "What can we remove from the interface this time around?"

That may sound like a weird question to ask - after all, a new version is about new features, isn't it, not removing them? Yes, but nobody said anything about removing features. The trick, the real trick, the difficult part that way too many companies do not work hard enough on, is "How can we add this new cool features while at the same time keep the interface as simple, or even simpler, than before."

A simple interface means an non-confusing interface and an app that is easy to use. It is not true that adding features and making the application perform more advanced functions does by necessity increase the complexity of the interface. It is only common. Common because especially with enterprise app development, the simplicity of the interface often is not considered important enough to merit much thought. Features are considered much more important, and there is an unfortunate tendency in certain circles to think that a cluttered interface communicates an advanced feature set.

We think a cluttered interface communicates laziness in the development team. A clean, simple interface that can communicate its own state, and is discoverable for advanced users, is paramount. For the users, and for us. If the users can easily figure out the interface, that means they don't have to ask us about it. Which means we spend less resources on support and more on developing the next cool feature.

You don't have to be an interface genius to make a great interface. It is enough to do like we have done, to constantly think about it, and iterate. And iterate and iterate. We will get there eventually.

I'd like to offer up surveyCanvas as an example of this. surveyCanvas is our module for producing the survey questionnaire. We had a really cool idea for this task from early on, making it into a true AJAXy, DOM-manipulating, OO JS, real-time, WYSIWYG... well, you get the idea. That was the concept, and that part turned out to work great.

Now, to insert a new element in the questionnaire, you'd just use the correct tool on the Tool panel. This is how the Tool panel area looked like in an early beta of surveyCanvas (it was then called SurveyDesk):

The page that contained surveyCanvas had a header with company graphics, and a logo. Very much web-page-like. At this point we were still struggling to get that genuine web application look and feel going.

As you can see, the Tools panel stretches out over the width of the page, and there are buttons for each function. Upper row, labeled 'Insert', contains all the different things you can insert in the questionnaire, and the second row contains commands for manipulating the same.

The thinking behind this layout was very much to have everything instantly accessible with the click of a button. So our concept of simple was more or less "fast". One click, boom!

Well, simple does not necessarily equal fast, of course. There are other considerations. There are already too many elements here, not to mention what will happen as you start adding more types of elements and more commands.

So, iterate.

Much better! We gathered all element types into a drop-down menu. Visually, we can now add as many more types as we like without cluttering up the interface. We've however had to trade in some operational speed for clarity. You now have to click to open the menu, click to select item type, then click to insert it.

The header is now just one thin grey line, and even though Options now has been tabbed with Branching, everything looks much cleaner.

Notice that we also now call items precisely that, instead of 'nodes' as in the beta. Calling what you add to your questionnaire a 'node' is something only a programmer would do, which is precisely why that term ended up in the interface....

So, now we are ready to add some features. Iterate!

Turns out the Options panel need all the vertical space it can get for certain question types, especially on small monitors. And it seems the Tools panel doesn't need so much space, so better to let the Options panel take over some of the real estate. Looks cleaner, too.

By this stage we feel we have firmly established that the Tools panel pertains to the elements you put down in your questionnaire, so the labels to the left are superfluous and can be omitted. The header is also completely gone, as in a desktop app, as the browser window itself shows the title.

There are clearly now more elements in the Tool panel than the last iteration, because more features have been added. We are not saying you should never add more buttons to a release, only that you owe it to your users and your support staff to work as hard as you can on simplifying the interface. Apart from the new tools, we wanted to direct the user to the very helpful video tutorial, so it gained its own button, too. This is how surveyCanvas looked as of version 1.2.

For the next version we are planning several new features in surveyCanvas. This again presents the problem of potential clutter in the Tools panel. What to do?

Indeed, what to do? A constraint we always have to work within is that of being a web application, i.e. running inside a browser. This is a challenge sometimes because: a) we do not have access to normal application features like a standardised menu system, and b) because many users have certain expectations about how things work on the web, expectation which might contradict expected behaviour in a normal desktop application.

For instance, the concept of the link is at the heart of the web. But we think of, and want to communicate, Laguna Survey as an application, so we do not use normal links inside the app itself, as we think that would dilute the concept (yes, links have been seen in desktop apps, but we have to work actively to be seen as a 'real' app). However, there are many places in Laguna Survey where a command is little more than a link, like the 'Reference manual' and 'Video tutorial' in the Tools panel in the illustration above. These then are presented as simple, custom buttons. We feel it is not appropriate to use the <button> element for these tasks, because we think many users have a mental model of executing a 'real' command with such buttons on the web. 'Sign in!'. 'Submit!'. 'Upload!'. The distinction is however far from clear-cut in all cases, and we are still struggling with this.

As an aside, let me tell you a little secret. You know that "Starting Laguna Survey..." message you see after you've hit 'Log in' and before you see the Laguna Survey main window? Its purpose? Getting all the user's data from the databases and set set them up in a way to then be used in the app? No.

Transferring the user's mental state from 'surfing the web' to 'using an application'?

Correct! We think the 'Starting' message has a real and important function, but it is not a technical one. It only takes a second to get all data on a user from the back-end servers and construct the main window. The 'Starting' message is pure psychology; when the main window has been loaded, the user must really believe she is running a real app, because that is what we are trying to serve. The 'Starting' message allows a mental transition, a mental wormhole. Leaving the web page browsing of the Sonora IT web pages and emerging in something that serves a web application.

This fact is not widely known outside of the development department, and in Laguna Survey's lifespan, everybody else has so far just taken the 'Starting' message for granted. Nobody ever asks about it. Not our users and not anybody internally. It works.

Back to the main thread. Another pain is not having access to application menus. A desktop application can 'hide' much of its richness and feature set in its menus, ready to be discovered by the power user. A menu on a web page, however, is something completely different. These drop-down menus were only ever meant to be for choosing options in a form. They were never meant for selecting commands to be executed, like application menus, so they behave accordingly: The first item is not treated as a headline, when you select something, the menu is stuck on what you've selected, etc.

So what to do? Well, if we add more commands, we could gather them in a drop-down menu with an 'Execute' button beside it. But that just doesn't make sense. If you choose one of the commands from the menu, it is always because you want to execute it, so why should you have to choose the command, then press a button? Why not just execute the command immediately? It is after all a command menu.

It might not be what the user expects from a menu on the web. But we have decided that is the way forward anyway. We have to take certain liberties if we want to keep innovating, and maybe users are ready to start expecting a different behaviour in web applications anyway? So, this is what the Tools panel area is going to look like in the next version:

The commands that you can perform on the elements are now contained in a single menu. Whenever you choose one of them, it is immediately executed, and the menu springs back to show 'Edit', just like an application menu would.

"But why didn't you make some custom menus, if the normal ones were not meant for that use, and users expect them to behave differently?".

Glad you asked! In short, the built-in menus have some properties that makes a lot of sense to take advantage of. They are standard, for one thing. They may look a bit different from browser to browser, but their behaviour is well understood and consistent. They display were they are supposed to, scroll if necessary, and are always large enough to display all contained text, horizontally. You do not have to cook up some custom JS and CSS to try to duplicate those aspects. But the most important feature is that the browser knows that they are menus. This is important in a wider perspective, in regards to for instance accessibility.

It struck us that the menu for inserting the elements should then also behave like that. You only ever use it to insert elements, so why should you click an 'Insert' button after you've chosen what to insert? You shouldn't have to, of course. So now, just choose the element to insert, and it is immediately inserted.

Now we have a cleaner, simpler interface with room for more item types and more commands, and we've reduced the number of clicks to execute them from three to two. We've added some more auxiliary commands and options but the Tools panel still looks cleaner than in previous versions.

And yes, we're going to add multiple undo/redo in the next version. We have it working in-house and it rocks!

Vi fikk et bra oppslag på Digi.no i forbindelse med lanseringen av versjon 2.0 av Laguna Survey. Vi er naturligvis glade for oppmerksomheten, og vi merket en god del trafikk fra Digi.no i etterkant. Det er dessuten hyggelig å se at den tekniske 'pressen' skriver om nye norske IT-firmaer, selv om de som oss har et kult produkt å vise fram men ikke så mye mer.

Det burde jo egentlig være nok, men det er ikke så lett å få oppmerksomhet. Det gjelder sikkert mange, men jeg har også en mistanke av at et produkt som Laguna Survey lett faller mellom flere stoler i forhold til å bli omtalt. Jeg mener, når leste du sist en test av flere forskjellige nettjenester først og fremst beregnet for nærligslivet? .... Det var det jeg tenkte.

Vi har iallfall til gode å se en eneste test, sammenliknende eller ikke, av netttjenester for å lage spørreundersøkelser. Utfordringen er hermed gitt.

We are very happy to announce that Laguna Survey 2.0 has now gone live on our servers!

As always, the new version is served to all new and existing customers.

There are many new features in this version, but the three major ones are:

- We are converting our price model into one where we charge for answers instead of number of particpants added to a survey. At the same time we are setting up all new surveys that you create with 25 (updated to 75 as of v.2.1) free answers.

We think this is a big deal for our customers because there is now nothing to pay if you don't get what is important to you: answers to your surveys. And even then, you get 25 (75 as of v.2.1) free answers for every survey. And even if you get more, it is up to you if you want to pay for some or all of those answers.

- Laguna Survey now supports open surveys. All new and existing surveys now have an open link associated with them that you can spread or publish on the web. When the link is clicked, the survey is taken. At the same time, explicit particpants will continue to work as before, and you use both the open link and explicit participants in one survey.

- Speaking of participants, you can now add participants in several batches, both before and after the survey has started. This means you now longer have to duplicate a survey if you for example get hold on a few more email adresses to people you would like to have as participants in a survey you have started.

We would additionally like to tell you about our new Laguna Survey User's Manual, which you can download from http://www.lagunasurvey.com/documentation

The user's manual takes you through the whole process of setting up and starting a survey, and more, step by step. It is heavily illustrated with screen shots from Laguna Survey.

Other new features and enhancements in Laguna Survey 2.0 include:

- The questionnaire is now editable in surveyCanvas even after survey start. surveyCanvas opens the questionnaire in a limited editing mode after the survey has started. In this mode, you can edit the text and graphics elements and certain visual setting.

- surveyCanvas now has a 'Print...' button in the Tool pane, which enables the questionnaire to be printed with 'nice' formatting, i.e. one page of the questionnaire will result in one page of the printout, and the graphical header, if you have inserted one, will show on each page.

- You no longer choose whether participants should be anonymous or not when you add them. Instead, you choose at the last possible moment, right before you start a survey.

Some error situations have been corrected:

- More stringent file checking has been put in place to files added as images in surveyCanvas, and as participant data in the Participants panel.

- The order of the asynchronous calls made to fetch the answer quantity and population in the Results panel has been corrected so that all numbers and percentages are always correctly displayed.

- A number of minor localization errors have been addresses.

- A problem with copying the questions from one survey to another has been corrected

- Several browser-specific issues have been addressed.

We are actually pretty impressed with ourselves and happy about version 2.0, and we really hope you are, too. Why can't you take a enterprise type task, make an application for it and program it to the brim with pro features, wrap it in a consumer-like friendly interface and sell it with a price model that appeals to big and small players alike?

We couldn't come up with a good answer to that question either.

The above illustration shows the image element selected in surveyCanvas, with an image inserted. You can use the File tool in the Options pane to insert and re-insert different images, which will replace the image already placed. The image inserted will be scaled down to 800 pixels horizontally, if necessary.

- New element type: Link. You can now insert a web link in the survey wherever you want, choosing whether the link should open in a new window or use the existing window (thus jumping out of the survey).

Markedsføringsstrategien vår for å reklamere for Laguna Survey dreier seg mye om online annonsering. Rent praktisk, siden vi er et forholdsvis nytt selskap, og nok i utgangspunktet har et noe skeptisk forhold hva folk i bransjen måtte finne på å si om strategier (og som de nettopp har publisert i bokform til kr 299, og ja, holder kurs om det neste uke til kun kr 4999, lunch inkludert), har vi brukt en del tid på eksperimentering.

Vi har bl.a. valgt ut tre kanaler for online annonser: Google AdWords, FaceBook Ads, og MyAd.no

MyAd har du kanskje ikke hørt om, men det er et norsk selskap med et konsept jeg liker: De plasserer annonsene dine på utvalgte nyhets- og bransjenettsteder. Du som annonsør har til en viss grad kontroll på hvilke nettsteder annonsen din skal bli vist på, gjennom et enkelt webgrensesnitt. Du kan plukke websteder direkte og/eller bruke nøkkelord (fra en før-definert liste) for å 'hinte 'om hva slags nettsteder som skal vise fram annonsen din.

FaceBook Ads vises på, nettopp, FaceBook. Du administrerer annonsene du lager i et webgrensesnitt, hvor du kan begrense framvisningen slik at annonsen bare vises for brukere som passer visse kriterier, f.eks. i en bestemt aldersgruppe, et bestemt land og/eller som har markert at de har et sett med visse interesser.

Googles AdWords system kjenner vel de fleste nettvandrere, iallfall fra brukersiden. Som annonsør 'laster' du annonsene dine tungt med nøkkelord. Annonsene vises da for dem som søker etter noe forbundet med ett eller flere av disse nøkkelordene, og bare da. Annonser vises både på Googles egne søkesider, og på myriader av andre websteder i forskjellige former, enten i forbindele med søk eller fordi innholdet på webstedet annonsen vises på er relatert til det du annonserer for.

Felles for alle disse tre er at du som annonsør kan betale pr klikk, dvs du betaler bare hver gang noen klikker på annonsen slik at de havner, i vårt tilfelle, på Sonora ITs websider. (Du kan i enkelte tjenester også velge å heller betale pr. framvisning, noe vi også har prøvd, men det ignorerer vi i denne diskusjonen).

MtyAd ble dessverre stemt ut først. Jeg liker som sagt idéen, og det skal baller til for å starte et online annonseselskap når man kjenner konkurransen. Men de harde fakta, når vi så på tallene, er at et klikk fra MyAd koster omtrent det dobbelte av det et klikk fra AdWords og FaceAds koster. 9-10 kroner vs. 4-5 kroner, for å kvantifisere. Skuffende, og overraskende at folkene bak tror de kan konkurrere med en slik pris. Det kan forsvares å ha en høyere pris hvis de klikkene vi fikk inn fra MyAd oftere førte til kontoopprettelse og tjenestekjøp. Vi har ikke sett noen slik effekt, og hvorfor skulle vi? Det er bare annonser framvist på nettsider med forholdsvis urelatert innhold. Det blir mer gjetning, men jeg lurer på om ikke folkene bak MyAd har gått ut og lovet dem som driver nettstedene de annonserer på en for høy inntjening, muligens rett og slett for å få innpass i konkurranse med andre ad-tjenester, og da selvfølgelig spesielt Google.

Da står vi igjen med AdWords og FaceAds. Som sagt fant vi ut at vi betalte omtrent det samme, 4-5 kroner, for et klikk fra begge disse. Vi hadde inntil nylig intet godt verktøy for å finne ut hvor kontolagere kom fra opprinnelig, men min mistanke var likevel at dem som kom fra et Googlesøk ville være i en modus hvor de mar mer villige til å opprette en konto. De søkte tross alt ofte aktivt etter noe i retning av Laguna Survey.

Den 6. mai gjorde vi derfor et grep for å teste dette. Vi satte alle våre FaceAds på pause, og så økte vi budsjettene på kampanjene våre i AdWords slik at forbruket på annonsering ble det samme. Vi overførte budjettet fra FaceBook til AdWords, om du vil.

Resultatet var slående. Grafen under viser hva som hendte med antall konti laget pr uke:

Trafikkvolumet inn til Sonora ITs websider forandret seg ikke signifikant etter dette grepet, men antall brukere som opprettet konto pr. uke fordoblet seg.

Nå er riktignok tidsperioden her for liten til å være statistisk vanntett, men effekten er uansett markant.

For å reformulere: vi bruker altså det samme beløpet på online annonsering som før, noe som betyr at vi får inn omtrent like mange klikk som før, men dem som kommer inn fra disse klikkene er generelt mye mer tilbøylige til å opprette konto. Brukerne som kom fra FaceBook må ha hatt svært liten sannsynlighet for å opprette konto.

Dette bekrefter den mistanken jeg hadde: Kommer du fra en Google-annonse som vises på et søk er du på jakt etter løsninger og derfor i den helt rette modusen i forhold til det vi er ut etter: at du oppretter konto og prøver ut Laguna Survey. Kommer du fra et klikk hvor annonsen vises fram fordi, vel, bare fordi den vises der av og til, er det veldig lite sannsynlig at du oppretter konto.

Sagt på en annen måte: har du søk i et ballegrep, slik Google har, følger det med et ballegrep på online ads.

Alt dette må naturligvis sees i lys av at vi er et selskap som selger en spesialisert webtjeneste. Landskapet kan se annerledes ut for ditt produkt.

Vi synes det utrolig kult at vi nå endelig tar Laguna Survey ut i det internasjonale marked, ved å lansere på engelsk på www.lagunasurvey.com og www.lagunasurvey.co.uk

Det har vært en god miks av koding og oversetting for å komme i havn. Vi begynte arbeidet med det tekniske rammeverket for omtrent et halvt år siden i skrivende stund.

Vi lanserer den engelske versjonen som Laguna Survey 1.3.1. I tillegg til at hele produksjonsapplikasjonen nå er tilgjengelig på engelsk, lanserer vi også som en del av denne versjonen støtte for å servere undersøkelser til deltagere på flere språk. Dette betyr at spørsmålsskjemaet deltagere får opp kan ha navigasjonsknapper og andre meldinger på norsk, engelsk, svensk, dansk, nederlandsk, tysk, fransk og italiensk.

Hvis du lurer på hvordan det tekniske oppsettet for multispråklige websider opererer i prakis, er ikke løsningen å lage kopier av alle websidene vi har på norsk og så erstatte all den norske teksten med engelsk, og legge disse sidene ut på en webserver som serverer til www.lagunasurvey.com. En slik løsning ville skapt en situasjon hvor kodeoppgraderinger måtte kjøres ut på to sett med sider, både de norske og de engelske. Man måtte også sette opp en server til. Dette er ikke en løsning, det er et mareritt. Og skal man publisere på flere språk senere må man ha enda ett sett med sider og servere, osv.

Løsningen vi bruker er å servere de samme sett med html- og kodefiler, fra den samme serveren, til besøkende uansett om de er på norsk eller engelsk. Et besøk begynner med at man kommer inn på vår server (det er i virkeligheten flere servere, men det bryr vi oss ikke om her) fra et eller annet sted. Hvis man har pekt browseren sin mot .com-varianten av vår adresse, plukkes akkurat det faktum opp av programvare på serveren og det blir satt en cookie hvor språkvariablen blir satt til engelsk ('en'). .com og .no domenenavnene våre peker til samme server, og kommer den besøkende til .no varianten settes språkvariablen til norsk ('no'). Slik kan programvaren hente de riktige tekststrengene for å sette inn på siden. Selve innsettingen skjer med en kodelinje:

Output( Localize('front_text2', -lang=$lang), -encodeNone);

Alle tekststrenger i systemet har en nøkkel tilknyttet seg. Nøkkelen er den samme mellom alle språk. Koden over sier: "Hent tekststrengen med nøkkel 'front_text2' fra spåket lagret i variablen 'lang' og sett inn her".

Utfordringen er lage en modul hvor det å arbeide med oversetting er lett, og som samtidig er hyperopimalisert for hastighet i det websidene skal bygges opp, av html, grafikk, data fra Laguna Survey og tekststrenger fra spåkmodulen. Admin-delen av Laguna Survey inneholder for eksempel over 150 slike tekststrenger, så det sier seg selv at å slå opp en tekststreng og sette den inn ikke kan ta lang tid.

I tillegg slipper man ikke unna noe tilpasset grafikk, typisk slik som inneholder tekst. Dette løses lett ved at de norske og engelske grafikksnuttene heter det samme men lagres på forskjellig sted, hvor navnet på lagringsplassen kan konstrueres opp ved bruk av 'no' eller 'en'.

Systemt er lettdrevet og ikke minst lett utvidbart med flere spåk.

Med versjon 1.2 følte vi at vi definitivt hadde fått Laguna Survey opp på et funksjonelt proft nivå. For å nå dit har vi lagt ned mye hardt arbeide med et sterkt fokus på det funksjonelle.

I et slikt hesblesende tempo kan det nok hende at utseende kommer som en ettertanke. Ikke i den forstand at vi ikke tenker design - en del av vår filosofi er at en bestemt funksjon må eksponeres på en bestemt måte i grensesnittet for å maksimalisere elmentets potensiale som verktøy. Måten det gjøres på, og dermed utseende på den biten av grensenittet, er en svært viktig bit av funksjonaliteten til det aspektet ved Laguna Survey.

Imidlertid har det nok ikke blitt så mye tid til å ta to skritt bakover og se på Laguna Survey som en designmessig helhet. Det har vi endelig tatt oss tid til. Resultatet har nå blitt publisert i form av Laguna Survey versjon 1.3.

Det viktigste vi har gjort er å skape en langt mer enhetlig og nedtonet fargepalett, og den har vi brukt gjennomgående i alle delene av Laguna Survey - hovedvinduet, hjelpevinduene, surveyCanvas og resultatvinduene.

Der hvor vi før brukte forskjellige toner av et ganske sterk grønt, vil du nå se en stålgrå farge med en svak teksturering. De forskjellige blåfargene er erstattet av én, i forskjellige grader av lyshet.

Meningen med denne forandringen er at Laguna Survey i mindre grad skal tilkalle oppmerksomhet mot seg selv, men at verktøyene inneholdt i Laguna Survey skal tre klarere fram, og dermed la deg gjøre jobben bedre.

Vi har videre rensket mer opp i utseende til surveyCanvas - Opsjonspanelet har fått mye mer plass og hele organiseringen av elementer er nå klarere og strammere. Vi har også introdusert noen animerte operasjoner for å gjøre visse operasjoner klarere visuelt.

Funksjonelt har vi rullet ut én stor nyhet i denne oppgraderingen (vi synes vi måtte ha noe funksjonelt nytt å by på også :-) :

Du kan nå splitte resultatdiagrammene på ett av de 'ett-svar' spørsmålene du har stilt i undersøkelsen. Du finner funksjonen øverst i Resultatdiagram-vinduet. Funksjonen er lett å bruke (én nedtrekksmeny) men litt vanskeligere å forklare formelt. Det er lettest å gi et eksempel: Hvis du f.eks. har spurt deltagerne dine om kjønn kan du nå splitte resultatdiagrammen slik at hvert spørsmål viser ett diagram med svarfordelingen blant menn og ett diagram med svarfordelingen blant kvinner. Et forsøk på en nærmere forklaring finner du i Referansemanualen for Laguna Survey på hjelpesidene.

Til sist har vi luket ut noen bugs: Problemer for IE7 å sette visse radioknapp-valg i surveyCanvas, browseren Operas problemer med å vise hode-grafikk i surveyCanvas og en obskur bug hvor resultatdiagrammene kunne vise feil data.

En ny versjon av Laguna Survey gikk ut i dag. Vanligvis er .01 oppgraderinger kun bug-fikser, mens .1 oppgraderinger inneholder nye funksjoner. I dette tilfelle har vi imidlertid gjort det litt annerledes. Denne versjonen er gått ut som et skritt på veien mot å lansere en internasjonal versjon på engelsk.

For å lansere en slik versjon er det flere biter som skal på plass enn bare språket. Versjonen vi slapp i dag implementerer følgende:

• Muligheten for å betale med kort i webbutikken. Fakturering er fremdeles mulig så lenge betalingsadressen har Norge som land og man betaler i norske kroner. Dermed er dette også presentert som valg i butikken. Vi har inngått en avtale med DIBS AS og PBS International for kortbetalingsløsningen. Dermed tas kortbetalingen i et sikkert grensenitt uten av vi trenger å håndtere kortdata.
• Man kan nå se priser i NOK, USD, EUR og GBP, og betale i samme, om man velger å betale med kort.
• Man kan nå, i surveyCanvas, velge at deltagerne skal se spørreskjema-navigasjon, feilmeldinger etc, på norsk eller engelsk. Flere språk kommer senere, på samme måte som hele Laguna Survey skal oversettes til flere språk senere.
• En siste forbedring som gjør at Laguna Survey er bedre i stand til å håndtere store kunder er at kjøpere av Business-abonnement nå selv kan legge til og administrere de brukerne denne type abonnement inkluderer. Dette gjøres enkelt fra 'Brukere'-menyen.

et par-tre mindre bugs har også blitt luket vekk, og vi har oppgradert jQuery javascript-biblioteket til versjon 1.3.1.

På denne måten får vi rullet ut deler av en internasjonal versjon, slik at vi kan konsentrere utviklingen mot kjernen av samme; rammeverket for å støtte flere språk, oversetting etc. Vi får samtidig erfaring med hvordan deler av dette oppfører seg i en produksjonssammenheng.

Vi følte samtidig at versjonen ikke fortjente en full .1 versjonsnummerering, ettersom de nye funksjonene er mindre deler av et større system, og spredd på forskjellige steder i systemet.

Digi.no, som refererer Finansavisen, melder i dag 12. februar, at Sesam har vært et enormt tapssluk for Schibsteds-konsernet, og har hittil kostet selskapet nær 500 millioner kroner, uten å i dag være verdt noe særlig.

Jeg husker godt at Sesam gikk ut med høy sigarføring i 2005 og erklærte at de skulle utkonkurrere Google på søk på det norske markedet.

Oh really? Hånd i været dem som så at det var en idiotisk idé med én gang!

I 2005 var Google allerede svært veletablerte og veldig store. Kunne Sesam ha klart det i 2002-3? Kanskje. I 2004? Neppe. 2005? Ikke nubbsjans.

I 2005 hadde jeg utviklet webstedet jobbfeber.no for Laterna Magica i noen år allerede. Webstedet var veletablert og godt kjent, og kom på topp i søkemotorene på visse kritiske søkeord som 'utdanning'. Vi hadde en betalt oppføring i kvasir.no, da (og fremdeles) en av de største katalogtjenestene i Norge. Her er en graf over hva som skjedde i løpet av 2005, hva gjelder hvor besøkende fant jobbfeber:

Ved årets inngang hadde Kvasir og Google ca. 90% av funnsteder/innklikksteder fordelt mellom seg, omtrent 50/50. Bildet endret seg dramtatisk i løpet av året, selv om vi fortsatte med den samme betalte oppføringen i Kvasir, og uten å foreta oss noe særskilt i forhold til Google.

Allikevel, Sesam-gjengen hadde stor tro på at de skulle kunne 'ta' Google ved å, så vidt jeg husker fra det de sa den gangen, skreddersy søkeresultater til norske forhold, språk og kilder. De skulle med andre ord være smartere med å tilpasse sine algoritmer til Norge. De hadde sikkert også ambisjoner om å få tak i innhold som ikke uten videre var tilgjengelig for andre søkemotorer, og gjøre slikt tilgjengelig på Sesam.

De fleste kan vel se at strategien har et fiaskopotensiale farlig nær den røde delen av skalaen.

En feilvurdering, som gjøres om igjen og om igjen og om igjen, er å lure seg selv til å tro at mens du implementerer den fantastiske programvaren med de smarte algoritmene, sitter konkurrenten din og tvinner tommeltotter og surfer videoer på The Onion. Med andre ord, du prøver å lage et bedre produkt enn det konkurrenten din har i dag, men glemmer å tenke på at når du lanserer 1.0 versjonen om 10 måneder, har din arge konkurrent, som har gjort dette her i mange år allerede, utviklet og sluppet to oppdateringer i mellomtiden. Og fordi du har måttet gå så høylydt og skrytende ut med det kommende produktet ditt, for å skape blest om det, har din konkurrent ikke hatt noen problemer med å forbedre sitt produkt akkurat slik at det du lanserer ikke har noen fordeler framfor kokurrentens i det hele tatt. Bummer!

Og trodde virkelig Sesam-folka at de kunne utkonkurrere Google på algoritme-nivå? Google er et selskap som ble startet av programmerere, som behandler programmerere som stjerner, som kultivererer programmeringsmiljøer på en måte som, vel, som har gjort Google til dem de er i dag. 'Nuff said, ingen grunn til å gni det inn.

Vel, så Sesam har måttet se seg om etter en ny strategi. Man kan jo ikke bare gi opp, ikke når man har Schibsted i ryggen. Så Sesam skal nå bli en katalogtjeneste. Lykke til med det. Så dere grafen over? Hva forteller den om folks vaner på nettet, hva gjelder å bruke slike tjenester? Hvor mange slike er det allerede i Norge? Eniro, Gule sider, 180.no, Nettkatalogen, Kvasir - det er dem jeg kommer på akkurat her og nå. Det er drøssevis andre. Tjener de penger? Aner ikke, men jeg har en følelse av at marginene ikke akkurat er formidable.

Men, de har iallfall skjønt én ting: Du kan bare droppe å konkurrere med Google på algoritmenivå. Du må gjøre det på konseptnivå. Google er først og fremst et globalt og generisk søke- og annonseselskap. Dermed må man prøve å være noe annet. Søk kan du glemme. Annonser? Du kan glemme å gjøre det på samme måte som Google, dvs på en generisk måte, hvor du selger ads som går inn i et gud-vet-hvor plasseringssystem på sider ved siden av søk. Hvorfor? Vel, jeg sa jo du kunne glemme søk. Bortsett fra det har jeg anektotisk bevis fra www.sonorait.no.

Når jeg ser på indekesering av sidene våre fra forskjellige søkeroboter, er Yahoo Slurp-bot'en en av de aller ivrigste. Vi ligger også høyt oppe på resultatlisten på www.yahoo.no på enkelte relevante søketermer, som f.eks. 'lage spørreundersøkelse'. Hvor mange innklikk fra yahoo.no får vi pga. dette? Nesten ingen. Helt neglisjerbart. Jeg synes nesten litt synd på bot'en til Yahho, som ivrig tusler rundt på sidene våre og suger til seg alt den kommer over, og så er det så få som bruker yahoo.no. Google dominerer totalt.

Så, hva står igjen i dette segmenetet?

Hørt om MyAd.no? Ikke? Antageligvis ikke. De er et nytt norsk online ad-selskap, og de har skjønt at de må gjøre noe annerledes. Det som tiltaler meg er at de jobber mot et bestemt antall nettsteder, slik at du selv kan bestemme på hvilke nettsteder annonsen din vises, dermed kan du profilere på en helt annen måte enn med f.eks. Google. Dessuten har de en enkel oppstartprosess og iallfall akkurat nå injiserer de litt virtuelle penger inn i kontoen din slik at du kan teste ut gratis. Kommer de til å lykkes? Aner ikke. Har jeg fått mye trafikk fra annonsen jeg kjører med dem? Nei, ikke ennå, men de er fremdeles i en betafase, så det er for tidlig å si. Jeg synes uansett de er på et veldig riktig spor.

Et annet ad-nettverk som har tatt en helt annen vei, og som gjør det bra, er The Deck. Jeg skal ikke forklare her, men ta en titt på deres website hvis du er klar for en annen måte å tenke på.

Som en julegave til oss selv har vi overhalt Sonora ITs grafiske uttrykk litt. Headeren i websiten har blitt lysere og ikke så dominerende, men først og fremst har vi tatt i bruk en ny logo for Sonora IT AS:

Alt grafisk design i Sonora IT har blitt gjort av undertegnede fram til nå, og vil fortsette å bli det inntil vi har en slik omsetning at vi kan ta oss råd til en skikkelig proff overhaling. Selv om jeg har gjort svært mye slik design opp gjennom årene er jeg på ingen måte proff. Resultatet er som regel iterasjoner av design, hvor hver ny versjon synes litt eller mye bedre enn den forrige. Så, etter en stund, begynner jeg å se problemene med designet og begynner å tenke på en ny versjon.

Den gamle logoen var designet rundt et stilisert fyrtårn, og var den siste i en rekke versjoner av dette temaet. Grafikken lengst til venstre under er den første publiserte versjonen, brukt i den tidligste betaversjonen av websiten, den lengst til høyre er den logoen som ble brukt inntil det siste skiftet:

       

Symbolikken ved fyrtårnet er vel forholdsvis åpenbar. I tillegg var inspirasjonen og modellen for fyrtårnet i logoen fyret på Jomfruland utenfor Kragerø, øya hvor jeg vokste opp. Litt nostalgi, rett og slett.

Problemene med logoen og et fyrtårn som logosymbol er imidlertid mange. Noen av dem er: Det er for konkret, det er for detaljert og gjør seg derfor ikke i små størrelser, og med to eller én lysstråle danner det enten en T-form eller en opp-ned L-form, noe som er vanskelig, iallfall for meg, å inkorporere i et logodesign som helst skal ha en definitiv bevegelse mot høyre/opp.

Den nye logoen henter sin inspirasjon fra firmanavnet, selv om det kanskje ikke er åpenbart. Da vi skulle opprette firmaet og lette etter et navn, var min første tanke at det burde henspille på den feedback man kunne få med produktene vi skulle lage, nemlig Laguna Survey og etterfølgende. Når man har vokst opp på en liten øy er det lett å assossiere det med en sonar. Der var imidlertid vanskelig å finne et firmanavn med "Sonar" i seg som var ledig, dessuten klinger det ikke egenlig så bra. Slik ble det Sonora  IT i steden. Sonora er et område i Mexico, og klinger iallfall bra i mine ører. IT ble lagt til fordi det faktisk er det vi driver med, informasjonsteknologi.

Men altså, sonar, feedback. Logoen viser både utovergående bølger og en 'radarsweep'. Samtidig former sistnevnte et utropstegn med sirkelen. Vi har beholdt blåfargen men samtidig introdusert fargekontrast. Logoen er tydelig også i små størrelser, ned til 16x16 piksler i adresselinjen i browseren (favicon).

Applikasjonsikonet for Laguna Survey har forøvrig gått gjennom enda flere iterasjoner, men det får bli en annen post.