One of our aims at Smart Insights is to make marketers aware of the many software tools that can be used to take advantage of the unique communications benefits of digital media.
Ultimately, these software tools should help customers too. Website personalization or behavioural targeting tools are an interesting class of tools since they can benefit both companies and customers, yet there are privacy concerns about how the data is collected is used.
This category of website personalization and behavioural targeting tools is also interesting since everyone still quotes the Amazon personalized recommendation "collaborative filtering" technique - read paper describing the algorithm. Most are still wowed by the accuracy of the technique (apart from when buying gifts for others...), but it doesn't seem to give rise to too many privacy complaints.
Magiq dynamic website personalization software
When I bumped into Malcolm Duckett at the last Econsultancy Masterclass presenting Magiq I was interested to know more. I know Malcolm from many meetings at the Emetrics Marketing Optimization summit where he was .
Here's my interview with Malcolm which presents the benefits and issues of managing this type of software.
Benefits of dynamic personlisation software
Q1 Please give some example of how Magiq is being used and the benefits it offers?
Two of our clients are using Magiq in quite different ways:
Hotel Reservation Service are using Magiq to build customer databases from the interactions of the visitor on their web site. Magiq records how users interact with the site and logs key information about their behavior. This includes information about their physical location (town/country), their organization, any search terms they have used (in reaching the site), other information which helps to characterize their interest (e.g. if they have registered for a newsletter, requested additional information etc.), and it also particularly looks for e-mail addresses that they have entered into the site.
Magiq makes this information available in two forms:
1 - as information available to Magiq's Gnomes, so that HRS can instruct the gnomes to present specific pieces of content for certain types of user. So for example if a customer has indicated that they are an SME or procurement manager, then the Gnomes can ensure that information relevant to them is injected into the pages presented. By using the Magiq historic customer profile they can ensure that the right information is provided on each visit, even if the user does not re-identify themselves.
2 - as customer lists. Magiq provides lists of people who are being very active (indicating that they may be "ready to buy") and would benefit from a call from their sales teams, and it also provides lists of people who look ready to defect (because they were regular users, but have stopped arriving). These lists are downloaded by HRS's sales and customer support teams to help them target their activities and communications to the right people with the right message. Because Magiq provides all the contact information for each visitor (that they provided to the web site) together with the data that identifies the customer segment they are in this communication can be very accurate.
FBTO were one of Magiq's early beta customers, they are an insurance company, and used Magiq's data to personalize the home page for different segments of users. They identified 4 stages of engagement in the brand, from users showing an interest in a specific product, to users who had purchased a specific product. For each segment and for each product group they created a small banner which provided specific information for users in that segment on the product group of interest.
By making this simple change they increased their conversion rates by 15%.
They also used the data to allow their call centre to contact people who looked to be on the point of buying, but had not made their purchase online. Targeting calls in this way produced a campaign that resulted in a sale to one in every three people they called.
Q2 What are the differences to other similar tools or the well known Amazon personalization?
Today there are behavioral targeting tools, that display content on the basis of current behavior, but we are unaware of any solution that builds and maintains long-term customer behavior records for every individual visitor, and lets the site owner use that combination of long term and real-time behavior to personalize their communications and web page content.
Also all other approaches to personalization or recommendation require the site owner to undertake significant tagging activities to collect the data needed, and to add tags to allow content to be added. Magiq by contrast is effectively Tag-Free, only requiring the inclusion of a static script file into the pages of the site. This can be done via a one-time change to the content management system or can even be done via the web servers or load balancers for the site.
This has several implications. Firstly, deployment of the solution is almost instantaneous, and secondly, a static/flat site or a site or application which the customer is unable to change can be turned into a site with dynamic content and rich media content using Magiq very simply. Moreover, the marketing team can change the content, personalization rules and content types instantly via Magiq's simple Apps and Gnomes.
In actuality, Magiq brings exactly the levels of personalization and creativity that Amazon has implemented, without needing to spend the millions of pounds in custom development and technology that they had to utilize to implement their solution.
In our view, Magiq is really Amazon personalization in a box!
Privacy issues of personalisation and behavioural targeting
Q3 How is the privacy of the user protected and how are possible concerns addressed?
Magiq only uses the data that the users voluntarily enter into the site and link this to location and organization data derived from their IP addresses....So it's like the message you hear when calling a call centre that says "this call is being recorded for the purpose of customer service quality improvement".
[Editor's note: Here is a grab from the Maqiq site showing some of the info collected]
Magiq's contracts also require the site owner to disclose the fact that they are collecting data and the purposes to which will be used, as we believe openness with the visitors is very important.
Magiq's solution also go to considerable lengths to manage the privacy and security of personal information being collected. For example, all data Magiq collects from the user is encrypted as it is captured so the transmission of any data more secure that the data normally passed form the Browser to the web site. We also provide opt-out functions for the user, allowing people to manage their own data collection and privacy decisions.
Lastly, unlike most "on-demand" services which are "multi-tenant", each Magiq customer has their own cloud instance - this means that each site's data is in a separate system the site owner has exclusive access and rights to the data collected. Moreover, the Magiq subscriber can decide if they would prefer their instance to be located in the EU or USA, thus easing data management and privacy compliance.
Q4 Which criteria should someone selecting a vendor of personalization technology consider.
Clearly this will involve the factors affecting the return on investment and cost of ownership on side and functional issues on the other.
Magiq aims to provide top-notch ROI by reducing deployment costs (via the Tag-Free deployment) and minimizing operating costs by using the latest cloud-computing technology and by allowing the user complete control over their expenditure via the Gnomes budgets. At the same time it is attempting to provide maximum return by providing usable/actionable data on the individual users and their interests, and by allowing one-to-one personalization (as opposed to the suck-it-and-see approaches provided by MVT tools and "herd-behavior" approaches of product recommendation solutions).
From a functional perspective each business needs to understand the functions they want to implement to maximize the effectiveness of their on-line marketing and communications. This will vary from business to business. Magiq's range of simple-to-use apps attempt to provide powerful personalization, optimization and CRM solutions on normal PC and mobile platforms. Other solutions focus on features like and product recommendations and simple AB testing which are functions not provided by Magiq today.
Also the roadmap of a supplier is important as investment in solutions in this area tend to be quite long-term (especially as the value and quality of customer data available grow).So it is important to choose suppliers with long-term commitment to the market, and a vision for the future.
So for example, Magiq provides a good platform based on individual customer data, and we plan to extend the family of apps to embrace other applications like email campaigns, dynamic pricing, cross channel data integration etc. and also extend it to embrace Flash, Flex, and standalone apps created with Adobe AIR.
Assessing the value of website personalisation tools
Q5 How would you recommend assessing the value of a tool like Magiq? Which KPIs should be reviewed.
Magiq is focused on the implementation side of on-line marketing, while it records all the activity it has undertaken it does not undertake the analytics part of the process, as most users already have Web Analytics or Customer Analytics and reporting tools, that we expect users to use to analyze the effectiveness of the campaigns and activities that Magiq implements.
The actual KPI's that should be measured and tracked are clearly customer specific. So for a transactional site these might be focused on revenue, and conversion rate. For non-transactional sites other metrics are important. Some customers will choose to focus on the visitor retention and engagement factors (duration of engagement, time on site, visit frequency and latency etc.), others will want to analyze the customer data provided by Prospect and Retain to assess the value of the audience that they have in monetary or other ways.