Avail Tech News

Personalized Emails made easy

2011-03-08T10:02:00.002+01:00

Avail has for years been used successfully in email campaigns, especially by our more early adopter customers. Recently we have noticed an influx in the interest for recommendations in emails and we decided to make it a lot easier for our customers to get Avail recommendations in emails.

In the Control Panel there is a new node called Email Recommendations. If you have the necessary data in place then getting email recommendations up and running is a matter of following an easy-to-use guide. The data necessary is product names, prices, image URL:s and product-detail-page URL:s. At the end of the guide you will have a HTML-code block that you simply copy and paste into the email service provider of choice.

A drop-down contains a list of major email service providers such as Responsys, MailChimp, Neolane and Apsis which will generate HTML-code that works out of the box. But as long as you know the tag-format, Avail email recommendations can be used with any email service provider that supports dynamic HTML.

Control Panel

2011-01-17T11:30:00.000+01:00

On the second of November 2010 we released the HTML Control Panel. After ironing out various kinks and fine-tuning the usability we will now make the HTML Control Panel the default Control Panel. As of February 28 2011 the JWS Control Panel will be disabled for all Avail SaaS customers and all Avail OnSite customers on version 8.5+.

One Method to Rule Them All - SaaS & OnSite G8.5

2010-11-23T14:26:00.000+01:00

With this release the method getRecommendations is introduced. This method will replace all current recommendation methods. Before this release the programmatic API method used defined what input could be used and the type for recommendation generated. With getRecommendations the input used and the type of recommendation generated is set in the Control Panel.

Why a new Method?
The first purpose is to empower non-techies such as e-commerce managers and marketing people to easily change the type of recommendation without having to rely on techies to do this. A safe bet is that this is something that will be appreciated by the person wanting the changes and techies alike.

The second purpose is to allow a more flexible set up of recommendations which gives merchandisers a larger degree of freedom to customize the recommendations to the need of a particular site.

The third purpose is to make it easier for new customers to integrate Avail and for existing customers that want to implement recommendations in new areas as there is only one recommendation method.

The forth purpose is to prepare for upcoming features that we are working on but not quite ready to announce.

Input

The new method getRecommendations takes multiple inputs. The available inputs to base recommendations on are:

Phrase - A search phrase used on the internal or an external search engine.
SessionId - The session id of the current session.
ProductId - A single product identifier. Typically the a product on a product detail page.
ProductIds - A set of product identifiers. For example a wish-list or a feed from the Facebook App My Bookshelf.
UserId - A user identifier of a logged-in or identified user.

When implementing getRecommendations make sure that you use all available inputs. This will allow non-technical people to readily change the input using the Control Panel.

Output

Screenshot

As mentioned the output will be controlled from the Control Panel. On the template level you will be able to set if the output will be products or people. If the chosen output is products, then you will be able to chose if the recommendations should be based on:

Visitors’ Clicks and Purchases
Visitors’ Searches
Manual Rules

On the subtemplate two new drop-downs will let you set the type of recommendation. Using the two drop-downs a sentence is created that we hope in everyday words explains the type of recommendation made. Example of sentences that are formed:

People who bought the products you have viewed also bought
People who searched on-site selected

Screenshot

Old Recommendation Methods
Old recommendation methods such as getProductPredictions and getRelatedActions will be deprecated. This means that they will continue to work for some time, but all new code with recommendation-requests should be done with getRecommendations. Eventually a transition has to be made as the old methods will stop working. No date for when the old recommendations methods will stop working has been set but this will be announced and communicated well ahead of time so that everyone can do the transition in a orderly manner.

Other Minor Features
With this release it will be possible to apply a dynamic parameter to all subtemplates in a template instead of just one subtemplate.

Also, for those using Avail to return presentation data the float output has been changed to always be at least two decimals. This will result in whole number or one decimal prices being presented as 25.00 and 36.70 instead of 25.0 and 36.7.

A new method has been added: logRemovedFromCart. This method removes a product added to cart with method logAddedToCart as stored in Avail. This is useful when doing recommendations with getRecommendations based on the content of the cart which is derived from the argument session id. It will also affect the products filtered with the template setting Exclude Added to Cart.

Avail OnSite G8.5
OnSite G8.5 will include getRecommendations and the features mentioned in the following posts:
More powerful Recommendations
HTML Control Panel

Code to handle availability problems

2010-11-22T12:05:00.004+01:00

We have devised a solution to handle the event of Avail SaaS not being available.
Before the Avail script tag add a row that declares a empty Emark object:
<script type="text/javascript">var Emark;</script>
Add an if-then statement before the Avail-specific code that checks if the object is available:
if(Emark){
//Avail Code
}

Full page sample:

<html>

<head>

function populatePredictions(){

if(Emark){

var emark = new Emark(false);

var recs = emark.getProductsPredictions(['0208931'], 'GETRELATED_DEFAULT');

emark.commit(function() {

document.getElementById("prodRecsDisplay").innerHTML += "Recs: " + recs;

});

}

</script>

</head>

<h3>Sample Page</h3>

<b>Recommendations</b>

</body>

</html>

More powerful Recommendations - SaaS

2010-11-02T17:00:00.001+01:00

Avail RnD has been hard at work and is now ready to release a new powerful algorithm for deriving recommendations. The new algorithm works on the click data collected from users. For now it will be released on Avail SaaS and on Avail OnSite in a couple of weeks.

Sequences

To improve the quality of the recommendations sequences are taken into account. A concrete example of when using sequences is beneficial this is that if you are contemplating buying a bike, a water bottle might appeal to you. It is less likely that if you are shopping for a water bottle that you might buy a bike on the spur of the moment.

Self Improving

The new algorithm is self learning and tweaks itself based on how users respond to the recommendations thus adjusting itself to vertical specific behavior. This is all automatic and you do not have do a thing.

Selecting Data Sets

To empower advanced users a new drop-down in the Control Panel lets you choose which data set you base the recommendations on.

Avail works with three data sets:

viewed - products clicked on during a visit
bought-together - products bought at the same time
bought - products bought by a user over time

To see how this could be of use, consider the recommendations based on the content of the shopping cart using the method getCartPredictions. The template has a maximum of three recommendations and two subtemplates:

Subtemplate 1 - returns 3 recommendations, no fallback
Subtemplate 2 - returns 3 recommendations based on the data set viewed with fallback recommendations

The effect of the above is that you will always have three recommendations. First Avail will generate as many real recommendations as possible based on the bought-together data set. If there was not enough recommendations, get recommendations from the viewed data set and if still not enough append fallback recommendations.

The HTML Control Panel Live - SaaS

2010-11-02T16:59:00.005+01:00

The HTML version of the Control Panel is live! The HTML Control Panel will work for all customers on Avail SaaS. In a couple of weeks we will release an Avail OnSite version that supports the HTML Control Panel.

To access the new Control Panel go to https://admin.avail.net. Starting this Thursday you will be prompted to use the HTML version when opening the old Java Web Start version of the Control Panel.

Supported browsers for the HTML Control Panel are Firefox 3.5+, Chrome 6+, IE 8+ and Safari 4+.

Metrics Area

Editing a subtemplate

In the new version we have tried to simplify several common tasks. Some examples of changes we have made:

In the Metrics areas you now have forwards and backwards buttons that allows moving the dates for the graph forward and backward in time much more easily than before
The "Get Graph" button has been removed and the graph is updated with every change you make
When hovering with the mouse over the graph the values of the different graphs are rendered as numbers
We have moved the test tab from the subtemplates up to a separate tab and redesigned the test area
When hovering with the mouse over a template name you can see an overview of the template settings without having to open the template
and much more...

Let us know what you think. Your opinion matters to us.

Henrik Schinzel

CTO Avail Intelligence

A new Control Panel on the way

2010-09-29T19:55:00.000+02:00

We are putting the finishing touches on a HTML version of the Control Panel. This will launch in a couple of weeks from now and will replace the current JavaWebStart Control Panel.

The Avail Control Panel been rewritten from the ground up in HTML and JavaScript. We did this as a few of our esteemed customer were not allowed to use Java on their machines due to IT policies. Also, the graphical interface of Java did not work consistently over all operating system creating a lot of extra work.

The new Control Panel will work with the upcoming Avail OnSite 8.4+ and Avail SaaS. All functionalities from the old version of the Control Panel will be available in the new version of the Control Panel.

When we launch the HTML version of the Control Panel you will be alerted of this fact when you open the JavaWebStart Control Panel and asked to start using the new version.

Enhanced Discovery & Upsell

2010-06-02T15:32:00.000+02:00

On Avail OnSite 8.3 and Avail SaaS it is now possible to use clicks as basis for recommendations. Using clicks enables better quality upsell and discovery recommendations for products that have been purchased few times. Examples of when products can have few orders are products with short life spans, newly released products and on newly launched sites.

Availability and Merchandising

2010-05-10T10:38:00.002+02:00

Two new features have been rolled out on Avail SaaS: the ability to update availability of products real-time and merchandising.

Updating the Availability

In Avail nomenclature a product available for being recommended is called valid. With the new API method setProductValidity it is possible to change the validity of a product real-time. You can change a product to being not-valid, which will cause it to not be recommended any more if you have set Only Valid to true in your template. Also, it is possible to make a non-valid product valid and thus making it eligible for being recommended. This method will only be available through the JSON-RPC interface.

Merchandising

It is now possible to do manual merchandising with Avail SaaS. Templates that previously have been used to filter the recommendation can now also be used to choose what products to display. Instead of a template filtering a set of recommendations, the template filters the whole product set. This gives you an very easy and flexible way to set which product to display on for example the start page, a landing page or a category page. Examples of what you can choose to display:

the five latest DVD releases
the most popular product that are on sale
explicitly set products, for example staff recommendation
top-sellers in a given category
and much more...

Which criteria that can be set up depends on the data uploaded to Avail SaaS. For example, to sort or filter on price, product prices data have to be uploaded.

In the Control Panel you will be able to use two new settings that work for merchandising only:

Column to Sort On - sets which product data column the result should be sorted on. Examples: price, release date, sales numbers.
Sort Order - sets if the sorting should be ascending (lowest first) or descending (highest first).

The new API method for invoking merchandising is called getProducts. The method supports dynamic parameters and returning product meta data just as the recommendations-methods.

Avail OnSite G8.2

2010-04-27T16:47:00.000+02:00

Today we have released a new version of Avail OnSite: G8.2. This contains a set of very strong features like:

A/B Testing
More detailed metrics
Improved recommendations
Multiple currency support
Auto categorize
and more...

For more information about the new features read the following two blog post from when the features were rolled out on Avail SaaS:

New Features On Avail SaaS

Metrics and Server Side API - SaaS

The only feature from these two posts that is not available is the server side API which is an Avail SaaS specific feature.

With these powerful new features we strongly recommend all customers to upgrade to G8.2 or to migrate to Avail SaaS. Let us know if you have any questions!

/Henrik Schinzel

Co-founder & CTO Avail Intelligence

Metrics and Server Side API - SaaS

2010-04-01T14:20:00.004+02:00

In approximately two weeks from now Avail will gradually begin rolling out new features on Avail SaaS. Approximately three weeks from now all customers will have received the update. The focus of the update is an overhaul of Metrics and a server side API.

Metrics

Metrics have been given an overhaul. The purpose of the redesign is to make Metrics more user friendly and transparent. Some measurements have received new names that should be easier to understand. The buttons have been moved around and grouped to make working with Metrics more streamlined.

Before logAddedToCart put a session-product only in the state added-to-cart and not clicked-on. Now logAddedToCart will put a session-product in both states, thus making the state added-to-cart a strict subset of clicked-on. This allowed removing some hard to understand measurements and giving others easier to understand names.

Grouping Templates

It is now possible to group templates into a single chart. As an example lets say that you have one template per main category on the product detail page and you want to see a unified chart for the performance of the recommendations on the product detail page. Then you select all the templates for the product detail page and check the Group Templates option which will result in a unified chart for the product detail page.

My Overview

My Overview is a new view that has been added. It is the first thing you see when you enter the Metrics area. The purpose is to give you a snapshot of the results over the selected time period.

Export Data

The Billing Info has been replaced with Export Data. This does an export of detailed metrics data. This data can readily be imported into Open Office, MS Excel and Google Spreadsheet for analysis.

Server Side API

A server side API has been added. Using JSON-RPC you can now invoke Avail from your servers. The purpose of this interface is to be able to use Avail for recommendations in order confirmation mails, call-centers, in-store kiosks and more. Due to the inherent latency of calls over the Internet, it is not advisable to use the server side API for high volume calls where low latency is critical, such as recommendations on web pages and newsletters.

Exclude Items in Shopping Cart

If this setting is checked and items added to cart is logged with logAddedToCart, items in a session's shopping cart will not be recommended to that session. To enable logging of non-recommendations added to cart, the TrackingCode argument has been made optional in logAddedToCart.

Do not hesitate to give us feedback on these new features and further suggestions.

/Henrik Schinzel

CTO Avail Intelligence

New Features On Avail SaaS

2010-02-10T08:44:00.007+01:00

We are very excited to announce a slew of important new features on Avail SaaS. We will roll out the new features to all SaaS customers over a period of two weeks starting this week.

There are three new features which we have named the "Big Ones" as we deem that these will bring large benefits for our customers. The Big Ones are:

A/B Testing - Will let you easily set up A/B tests of Avail templates to find optimal settings.
Template Tracking - Allows you to track a lot of metrics on individual templates.
Improved Recommendations - An addition to the algorithm improves the quality of the recommendations.

I hope you enjoy the fruits of our hard labor!
/Henrik Schinzel
Co-founder & CTO Avail Intelligence

A/B Testing - Big One 1

A/B Testing will allow you to easily and rapidly set up A/B tests of templates. The purpose of A/B testing templates is to optimize the settings used. Example of tests that you might want to do:

should the price rule on the shopping cart template state that the price should be less than 10, 15 or 20?
should you should display three or four recommendations on the product detail page?
should the third recommendation shown be a product on sale or a new release?

You can

Create up to five variations of a template
Assign a percentage to the variations that sets how much traffic the variation gets
Set start and end date-time
Edit the variations like a normal template

When the test finishes a report with the results of the tests will be generated that you can access in the Control Panel, and if you have set an e-mail address the report will be sent to you.

Some of the metrics available in the tests:

Sales
Conversion Rates
Average price purchased recommendation
Number of displays

Implementation Required
Requires Template Tracking
Automatic

Template Tracking - Big One 2
Several customers have expressed the desire to have detailed metrics in Avail. We have heard you and are happy to announce Template Tracking. This will allow tracking individual templates on metrics such as

Sales
Number of clicks over number of views
Number of views
Average price of recommendation purchased from a template
and more ...

The data will be visible in the Metrics area

with a lot of metrics to choose from

Implementation Required
The implementation required - assuming that metrics is already used - is to use the new methods:

logClickedOn(ProductId, TrackingCode)
logAddedToCart(ProductId, TrackingCode)

when a recommendation is clicked on or added to cart. The TrackingCode is returned with the recommendations.

Improved Recommendations - Big One 3
The quality of the recommendations has been be improved. To what degree will vary from customer to customer. What has been done is a more effective noise reduction. Noise is data that makes it more difficult for a computer to correctly classify something. If you have a product that has sold really well like the book The Da Vinci Code by Dan Brown this has noisy data and causes noise as it is connected to so many different people and products. The new algorithm more effectively reduces the noise from data that is not as relevant and therefore is able to produce better recommendations.

Implementation Required
Automatic

Summary of the Big Ones
To recap the three new important features:

A/B Testing - Will let you easily set up A/B tests of Avail templates to find optimal settings.
Template Tracking - Allows you to track a lot of metrics on individual templates.
Improved Recommendations - An addition to the algorithm improves the quality of the recommendations.

Auto Categorize
Auto Categorize solves the problem of categorizing the recommendations and more importantly the fallback recommendations. If category filters are not used with recommendations requests on a product detail page and there are no real recommendations the fallback recommendations are a plain top-list with no relationship to the product currently viewed. For example, if viewing a goth metal album that no one has bought and no categories filters are used, the recommendations will be from the top list and typically be products like albums from Britney Spears and Justin Timberlake which will most likely not appeal to the goth metal crowd.

Auto Categorize applies to recommendations based on a single product, a set of products or a user. If used, the returned recommendations and fallback recommendations will automatically be submitted to a category filtering based on the categories of the input. In the example above, if the goth metal album belongs to the categories Goth and CD the returned recommendations will be from the top-list but within the categories Goth AND CD and thus much more relevant.

Logic
If the input is one product

The returned recommendations will belong to all of the categories of the argument product.
An AND-relationship between the categories of the argument product.

If the input is several products

The returned recommendations will belong to all of the categories of argument product 1 or all of the categories of argument product 2 and so on.
An AND-relationship between the categories of each argument product and an OR-relationship between these.

If the input is a user

The logic of several product is applied to the purchases of the argument user.

Implementation Required
Automatic

Cart Recommendations
To optimize the quality of the shopping cart recommendations Avail now supports cart recommendations that is intended to catch cart specific behavior such as recommending a toner with a printer.

Implementation Required
When uploading transaction history there is a new optional column, order id. There is a new method getCartPredictions that returns shopping cart recommendations. To incrementally add more order data the method logPurchase now accepts an OrderId as optional argument.

Maximum Number of Recommendations from a Template
With the new release it will be possible to set the maximum number of recommendations a template returns no matter how many recommendations the subtemplates are set to return.

The reason for this feature is that in some implementations of Avail a limitation on the number of recommendations has been hard coded in the web layer. This hampers the merchandiser in that he or she cannot readily change the number of products that are returned from a template using the Control Panel. To change the number of recommendations displayed the merchandiser needs to request resources from the techies.
By adding this feature there is no need to implement a hard coded limitation on the number of recommendations rendered in the web layer. The merchandiser can control the number of recommendations returned without bothering the techies. Also, by adding this feature it is one less thing for the techies to do when implementing Avail if this behavior is desired.

Complex Templates
This feature also enables more powerful merchandising. For example, we can create a template that is set to return a maximum of three recommendations that has three subtemplates:

Subtemplate 1 - returns three recommendations with Price < 5, no fallback recommendations
Subtemplate 2 - returns three recommendation with Price < 10, no fallback recommendations
Subtemplate 3 - returns three recommendations with Price < 15, use fallback recommendations

The result is a template that generates three recommendations that are as cheap as possible. This as achieved as the first subtemplate generates as many real recommendations as possible that cost less than 5. If there is not enough real recommendations that cost less than 5, real recommendation that cost less than 10 will be added. And lastly if it will add real recommendations that cost less than 15 and use fallback if necessary. This is just one example with what you can do.

Implementation Required
Automatic

Category Names
Avail now supports category names. If category names are uploaded these will be displayed in the Control Panel instead of category identifiers.

Implementation Required
New upload: CategoryNames

Find My Twinsumer
A new method accepts a user identifier as argument and returns the user with the most similar behavioral pattern. This feature does not stand on its own but is a powerful feature to any e-commerce website, because it allows visitors to the site to quickly identify "people like me" (anonymously) and then implicitly ask these "Twinsumers" for product-related information and recommendations.
The way a retailer can use this function is e.g. to publish the following functionality on any of their website pages. "Find My Twinsumers (people like me), and then:

get a list of products these Twinsumers have purchased (optional: in this category)
get a list of products these Twinsumers have written a review about (optional: in this category)
get a list of products these Twinsumers have given a rating higher than X (optional: in this category)
get a list of products these Twinsumers would probably also have bought, had they come to our site today (optional: in this category)
let me ask them a question (via anonymous email/messages that the site can distribute without disclosing personal email addresses)
let me send them "invites" to any community of choice like Facebook, Twitter, LinkedIn, etc (via anonymous email the retailer distributes)

Implementation Required
New method: getTwinsumers

Clickstream Tracking & Recommendations
Avail now automatically collects data on which products visitors have clicked on and enables recommendations on this clickstream data. The products clicked on by a user is collected with getTemplateClickstreamPredictions and logClickedOn. The purpose of this feature is to enable more rapid integrations.

Implementation Required
Requires Template Tracking.
New method: getClickstreamPredictions

Multiple Currencies
Metrics now has multi currency support. With this it is possible to log purchases in different currencies. The purpose is to facilitate integrations. Note, there is no automatic currency conversion.

Implementation Required
New optional argument in logPurchase: logPurchase(..., [Currency], ...)

Expanded Identifiers (AlphNumString)
The data type for user, product and category identifiers (AlphNumString) now supports three more characters:

at sign (@)
full stop (.)
hyphen-minus (-)

Implementation Required
Automatic

Recommendation Reinforcement - Beta
This is a beta feature which is disabled by default and will be evaluated over next few months.
Some products trigger customers to buy more than others. The purpose of this feature is to give a bump to the prediction score for products that have a high inherent conversion rate. This will make these high converters more likely to be recommended.
This is especially useful for products that are not so easily found. Some products are easily found because:

they are prompted - on the first page, on category landing pages, on banners and so on
rank high in the search results
are famous on their own accord such as the iPod

There typically exists a set of products that are high-converters that has not been caught by the merchandisers and thus are not prompted. But when a user stumbles over these high-converters these are more likely to trigger a sale than the average product. By giving these high-converters a bump in the predictions score, Avail will automatically promote them in the recommendations.
Logic
The increase of the prediction score is the ratio: the number of purchases to the number of clicks. The maximum increase in prediction score is 20%. This logic is likely to change over time.

Implementation Required
Requires Template Tracking
Automatic

Landing Page Recommendations in Avail SaaS

2009-07-10T09:43:00.004+02:00

Avail SaaS has a new feature, landing page recommendations. This allows e-commerce sites to create a dynamic landing page that will display products based on the search phrase used on an external search engine such as Bing or Google.
Avail will continue to roll out functionalities from the eMarketing Suite onto Avail SaaS.

Maintenance Release

2009-06-29T13:45:00.004+02:00

Avail has just released a maintenance release, G8.0.31. It is an incremental upgrade to G8.0.21 that was released at the end of last year. This release contains a set of bug fixes and set of minor new features. The upgrade is available free of charge to all existing customer using the eMarketing Suite. Contact support[at]avail.net for more information.

Important updates:

The Microsoft drivers have been deprecated as these cause JRE crashes and memory leaks. JTDS drivers have been added instead.
New Logger that logs to standard out. Very useful during debugging and development.
CIB is now more memory efficient and has improved multithreading.
FileLogger now has rolling seven day file log.
Unpublic reviews do not contribute to average rating anymore.
New command added addBillingReport. This sets up a monthly report that sends billing info to Avail.

Important Bug fixes:

Float arguments were incorrectly set to 0 before.
The Web Prompt has been updated to handle encodings correctly in IE and results are no longer cached in IE.
NTEventLog now works when running on a 64 bit JRE.
Before uploadCibData always reset protest limit to default value -1.
Windows service max memory setting was previously ignored.

For a full list contact Avail Intelligence.

Avail SaaS launched

2009-05-12T19:12:00.015+02:00

Avail is proud to announce that Avail SaaS has launched. Starting now Avail offers recommendations as a software as a service.

For those of you who are acquainted with the eMarketing Suite we have launched the functionalities of the Navigation Predictor and the Social Search Optimizer first. More modules will be added in the near future.

Recommendations are requested client-side on the browser using JavaScript in a solution similar to Google Analytics and other well known SaaS services. The recommendations are administrated using the Control Panel in the same manor as the local install version of the eMarketing Suite.

Check out the details at www.avail.net/saas

Avail SaaS is hosted on Amazon Web Services (AWS) for reliability and scalability. Hosting on Amazon offers the possibility to add a large set of machines to our cluster within minutes. Also, the solution is hosted at two physically separated data centers for reliability.

AWS is what is referred to as a Platform as a Service (PaaS). Developing on a PaaS is fantastic, especially if you have experience with setting up physical servers. No more power outages, maxed out power, cooling that stops working, redrawing of cables, following unlabeled cables and so on and so on. At first it was a bit of a threshold and there still are some challenges, but once past this the flexibility offered with a PaaS is fantastic. With PaaS hardware architecture becomes as easy to remodel as software architecture. Also, it affects software design as it adds a new axis of freedom - CPU. Before CPU was a fairly static maximum that was core in the design. Now with a PaaS one can start up 20 powerful machines for an hour to crunch some numbers and then shut them down. For example, an e-commerce site typically has a peak after lunch and after work, in between there is a lot of unused CPU. With a PaaS machines can be added to handle the peak loads and then scaled down to the number of machines that are appropriate for load. There have been services similar to AWS before, but nothing as easy and flexible as AWS. Noteworthy competitors to AWS are MS Azure and Google App Engine.

eMarketing Suite G8

2008-12-04T07:42:00.009+01:00

The eMarketing Suite G8 was released 2008-11-21 and is a technical release with several things rebuilt and a set of new technical features.

Navigation Predictor

To accommodate the ever larger customers that chose the eMarketing Suite the Navigation Predictor has been rebuilt from scratch. The data uploads are now 5 times faster and it uses one fifth of the RAM. This is on a data set of 5 million transactions using the file retriever.

Customer Interaction Broker

The CIB has a new entity, Questions and Answers. The Q&A allows any user to post questions about a product and other users to answer the questions. A user can ask a maximum of one question per product and a user can only answer a specific question once. Also new is:

safe transactional saving of CIB data
arbitrary top-lists on any attribute
add any property such as image, description and so on to users
get all reviews, lists, questions and answers created the last x days
data is stored more efficiently
review and list title are not also subjected to the profanity filter

Social Search Optimizer and Landing Page Optimizer

The SSO and the LPO now allows setting a minimum number of times a search phrase or an adword has to be related to a product for it to qualify as a recommendation.

Collaborative Searcher

CS now supports dynamic alphabets. Prior to G8 the CS was limited to the English Alphabet. Now alphabets of any size and any characters can be supported.

Administration

The Administration Interfaces (AWS and PHP) have been replaced by the Java based Control Panel.

The Control Panel is designed for multiple instance administration and is available in English, German and French. It automatically updates itself as new versions become available.

New Metric

The metric tool built into the eMarketing Suite has been rebuilt. The old A/B testing was not accurate on smaller data sets.

Message Handling System

The Message Handling System got a lot of attention in the G8 development with focus on logging and encoding.

There are now four types of loggers:

File logger - logs to file
NT Event Log - logs to the Windows Event Log
Syslog - logs to the sys log on Unix and Linux
MailLog - logs to mail

Any message handler - listener, command file loader, task, ... - can have any number of error loggers and a maximum of one event logger.

A new message handler has been added, Reports. This enables any eMarketing command to be executed as a schedule task and the response sent as a HTML-mail.

The AWS web server has been removed. Use the AJAX server instead which has been rebuilt on top of the popular open source web server Jetty.

A help command has been added which allows you to get help in any interface on any method. The help command supports wild cards.

Nuts and Bolts

The directory structure of the eMarketing Suite has been cleaned up and restructured to better support multiple instance set-ups.

All data uploads are now completely seamless.

The file retriever now support delimiters of any unicode character or any length. It lets you manually define which encoding to use, supports BOMs and supports uploading from gzip files.

The eMarketing Suite can now be set up as a Windows service out of the box.

A new tool has been added, Heartbeat. This tool monitors one or several instances and sends an e-mail if an instance stops responding and another e-mail when and if the instance goes up again. The Hearbeat tool also has a web page that gives an overview of the monitored instances. The Hearbeat is independent of the eMarketing Suite and should preferably be placed on a separate machine.

eMarketing Suite G7.3

2008-04-11T08:14:00.008+02:00

The eMarketing Suite G7.3 was released 2008-04-07 and is a release focused on the Collaborative Searcher.
New in Collaborative Searcher
Dynamic Score Thresholds
We heard a loud and clear feedback from our customers that the Collaborative Searcher returned too many search results, i.e. it was too fuzzy. To remedy this dynamic score thresholds have been introduced.

With eMark G7.3 the Collaborative Searcher will have three score thresholds:

Interval03 (Default) - Based on the search points for the whole result set, this threshold limits the results to those that are in the top two-thirds of the result point range.
Interval05 - Based on the search points for the whole result set, this threshold limits the results to those that are in the top half of the result point range.
None - Returns all results.

As all users of the Collaborative Searcher were of the opinion that too many search results were returned we made the dynamic score threshold Interval03 the default score threshold.

Sub Sorting on Search Results Score
Two new features in G7.3 can be combined in a very useful way. The two new features are:

A new feature in G7.3 is the use of sub sorting with the argument ColumnToSortOn - see below for details.
All search-methods now return abstract action data column consisting of search result score named _score.

These two features combined allow the search results to be sub sorted on an arbitrary action data column.

For example; if a user searches for "simpsons" and there are more than a hundred movies called simpsons that are all perfect matches, it is now possible to sort these on search result score and sub sort on for example sales last month. This would give the search results in order of relevance, but with the top-sellers first when two or more products have the same search result score.

Default Number of Cluster Words
The default number of cluster words returned from getPhraseClusters has been changed to seven. Before the method returned as many as it had.

Bucket and Permutations Deprecated
The settings buckets and permutations have been deprecated. It was evident that the settings were too complicated for people to use without extensive explanations. As Avail try to keep concepts simple we removed these settings. The settings will still work but have been removed from documentation and status messages.

Other New Features
Java Client
The Java client has been made Java 1.4+ compatible. Before the Java client required Java 1.5+. But as many Java-web-servers run on Java 1.4 we made it compatible with one version lower.

Sub Sorting
The argument ColumnToSortOn - used with Collaborative Searcher and Merchandising - now supports sub sorts. This means that you can sort on multiple action data columns. For example, if you have uploaded price and margin as action data you can now ask for returned data to be sorted on first price and then margin. This means that the returned results are sorted on price and any products that have the same price would then be sorted on margin.

The syntax for sub sorting is:
[+/-]ColumnName1,...,[+/-]ColumnNameN
If a plus or minus is present before a column name this overrides the argument SortAscending. + signifies sort ascending and - signifies sort descending.
Example:
+Tile,-ReleaseDate

Dynamic Parameters
With a release G7.2 just before Christmas an easier to use and more powerful version of dynamic parameters were released. The old version of dynamic parameters still works but has been removed from the documentation. The new syntax is in a more transparent human readable form and allows setting settings on individual subtemplates.
The new syntax is:
[append/replace] [andcat/orcat/notcat/rule] in subtemplate [subtemplate number] with [data]
Examples:
append rule in subtemplate 2 with Price<10
replace notcat in subtemplate 1 with Hardback

The data position for categories can accept multiple categories separated with comma.
Example:
replace andcat in subtemplate 3 with Drama,DVD,OnSale

New Site: Avail Tech

2008-04-07T15:46:00.004+02:00

Avail launches a new site - Avail Tech - which can be reached at www.avail.net/tech. This is intended for customers of Avail and contains up to date information about the eMarketing Suite.

The site contains:

The FAQ that previously was located on the main site
A blog
The complete eMarketing Suite documentation
Online training videos on the Administration Interface
A demo site that shows the different features of the eMarketing Suite
Feature request form in which you can submit feature requests and product suggestions

Avail Tech requires a login which you can get by mailing support[at]avail.net. It is one login per company.

G7.2 - Merchandising

2007-12-07T23:02:00.000+01:00

As an early Christmas present the Merchandising module was released in December of 2007. Before the release of this module the templates were used to filter the results from the other modules. Now the templates can be used to filter the whole product set. Thus, using the administration interface templates can be created that show the products you want. For example, the five products within the category DVD sorted on release date.

G7.1 - Metrics

2007-10-04T22:53:00.000+02:00

In the marketing material Avail boasts of sales increases. The customers that have measured the sales enhancing benefits of the eMarketing Suite know that the claims made are valid. Unfortunately not all our customers have had the time or the tools to implement this measurement. To fix this Avail in the fall of 2007 released the Metrics module. To make sure that as many as possible are aware of the effects of the eMarketing Suite we decided to make this module free for all Avail's customers to use. The Metrics module performs an automatic A/B-test. When a visitor enters the site he or she is randomly assigned to either the A- or B-group. Default one out of a hundred is assigned to the B-group.Throughout the session the A-group is shown normal eMarketing results and the B-group is shown blank or randomly selected top sellers. The Metrics module can measure:

Average Order Value
Conversion Rate
Sales

The results can be accessed and inspected in the administration interface.

G7.0 - Collaborative Searcher™

2007-03-15T22:38:00.001+01:00

In the spring of 2007 the Collaborative Searcher™ (CS) was launched. The CS is a fast fuzzy search engine that uses social intelligence to optimize the search results. Focus for the CS - like all Avail's modules - was performance, simplicity and social intelligence. Our aims were:

to create a search engine with outstanding performance
make the implementation and administration as easy as possible with no or minimum impact on existing architecture
use the social intelligence to optimize the search results

Today the CS runs live on Ellos (part of Redcats Nordic), DVD.co.uk and Home Entertainment (part of Bonniers).

In essence the CS does three things; searches, finds related words and collaborative searches.

1 Searches
The CS has a normal text logical search. This handles fuzzy searches and has very good performance. The reindexing of a large dataset (more than 2 million products) takes less than 5 minutes. The CS uses the same filtering system as the rest of the modules. This means that the searches can be sorted on any data you choose to upload and support multiple sort orders which enables sorting on the English, German, Swedish, French and so on alphabets. The search results can be filtered on any combinations of categories. The rule system allows creation of rules filters. For example, return only search results that cost less than 10 Euros and have been released the last year.

2 Finds clusters of related words
Using social intelligence the CS can learn which words belong to each other according to the view of the customers. For example, the word "Apple" has three meanings to users on a book site.

For most people Apple means a computer or the company. The related words in this cluster were mac, macintosh, g5, jobs and so on.
The second most likely meaning was the fruit. The related words in the cluster were gardening, tree, pruning and so on.
At a distant third meaning was the meaning the record label with only one word in the cluster; Beatles.

Making the related words clickable and the click leading to a new search including the clicked on word, yields a search navigation system which the customer can use to find the product he or she is looking for. This is to the best of our knowledge a brand new way of navigating a site. We know that Google has a similar functionality but it does not work in the same way and we beat them to it.

3 Collaborative Searches
The collaborative searches are social intelligence searches. Instead of searching just on the search phrase submitted by the user, the search engine automatically adds the related words from the first cluster to the search. For example, if a user searches on just "harry" on a book site today we know that the user is looking for Harry Potter books in general and specifically the first and seventh book. It is difficult or impossible for a purely text-logical search engine to understand that this is what the users are looking for and return these books among the first results. The Collaborative Searcher™ would append the related words from the first cluster transforming the search on "harry" to a search on "harry potter rowling hallows deathly stone sorcerer prince blood" thus returning the most relevant results first.

A common headache for people that have been administrating a search engine is the list of synonyms. The related words mechanism is a replacement for synonyms. The related words are how customers perceive that words are related to each other in the closed world of the site.