opinion
For the internet to continue to be relevant in our fast pace of life it must continue to evolve and adapt itself. Father of the web Sir Tim Berners-Lee has even begun hypothesising what the next stage of the internet's evolution may be; specifically what 'Web 3.0' will deliver. Far from being another step in the user generated content vein of Web 2.0, Berners-Lee has opted for something far more radical, and potentially even more revolutionary.
Under the guise of Web 3.0 a 'semantic web' (semantic defined as 'meaning') would see intelligent decision engines sifting through the vast quantities of information online, refining it into a more succinct package for human users. Sophisticated decision engines could not only find and retrieve relevant information (as search engines already do), but could combine it to form fresh understanding, delivering conclusions and recommendations depending on the information found.
When searching for and purchasing a holiday the engine could conceivably provide back to us all of the related information that we might otherwise have to seek separately - for instance checking the weather for the dates we're booking, recommending the most appropriate clothing to take whilst checking local events, storing the dates in our diary and calculating the amount of currency we should take with us. The possibilities for such artificial intelligence are truly endless.
However, for this evolution to occur, a giant leap is needed in the sophistication of the decision engines we create. Without the reasoning present in the human brain, how could a machine possibly make the subtle yet fundamental distinction between fact and fiction, or even be able to access the jumbled mess of information the internet is made from? Suddenly Web 3.0 is more science fiction than science fact.
Towards tomorrow today
Whilst a true semantic web seems years or even decades away, the basic principle under discussion - namely the use of decision engines in combining data to create a meaningful outcome - is much more achievable.
Direct marketers have been using data to make decisions for years. Whether using profiling information to create segmentations or embedding analytical models in campaign selections for individuals, skilled data analysts are a key element of the DM process. The speed of digital transactions has however moved this decision from planning cycles of months to minutes. Digital channels allow consumers and brands to interact with one another in real-time. Instead of the traditional marketing route of analysing customer data, targeting specific individuals and contacting them via the post or phone (which may not take place until weeks later), the opportunity now exists to make relevant communications at a very specific and immediate point in time. In parallel with this reduction in timescale has been the realisation that effective personalisation of communications requires customer data reflecting both online and offline activity. Accessing this data and processing it at digital speed moves us towards the semantic web
Whilst achieving the semantic web is blocked by the unstructured nature of the internet, marketers benefit from the data they collect and hold being well organised and easily accessible by human or technological approach; marketing databases are some of the most sophisticated in existence, yet built with easy and efficient data retrieval front of mind. Aligning this data with skilled analysis to create real-time decisions will create the personalised experience to which marketers aspire - we call it a 'positive interaction'.
For the consumer, this might be achieved when being dealt with in a way that they find straightforward, being given good recommendations based on their circumstances or getting exactly the right product they want. For the brand a positive interaction could be a sale, an increased likelihood of the customer returning, or even their advocacy about the brand to other consumers.
Digital channels lend themselves to achieving positive interactions, improving on the relevance and timeliness of customer communications, by enabling real-time interactions to take place. The wealth of real-time data generated by these channels is also a key precursor of their success.
Real-time interactions
Whilst most traditional direct marketing uses offline data only, online marketing is beginning to reflect online behaviour - some brands offer specific web content to consumers based on what pages they are viewing, or where they have arrived at the site from (clicking through an online ad from a particular destination).
The real opportunity for brands that desire a relationship with their customer is to amalgamate key data from both offline and online interactions and transactions. For example, using online data (such as what a consumer is looking at on the site) in combination with known offline data (such as their purchasing history) to generate web content, an email or a mobile message that offers information relevant to the customer at that time. Using real customer knowledge to drive the personal customer experience is a great example of a positive interaction - for example, if the customer hasn't made a purchase for several months or perhaps often gets as far as adding something to their basket but not completing the purchase, an additional incentive could also be included to help persuade them to buy.
To achieve this, real-time data generated online needs to be married, in real-time, with data stored and collected offline and used with analytics and decision rules, also in real-time, to determine the next message and deliver it, resulting in an interaction with the consumer that is precisely relevant and timely.
The 'digital brain' approach
It is here that the development of a sophisticated decision engine with the ability to combine existing offline data with real-time generated online data begins to illustrate the potential of a semantic approach. Dubbed a form of 'digital brain', this concept is designed to mirror the human brain in its ability to consider a variety of data sources when making the decision as to the best way to interact with the consumer at that specific moment.
Firstly the digital brain identifies the customer who is generating fresh behavioural data, perhaps interacting with the brand's website or telephone agent. Generally with the customer's permission, it then mines the offline database to access other information available on the customer, such as measures of their purchasing history, demographic information and channel preferences. Algorithms developed through analytical insight are used to link these data sources together and deliver a decision, resulting in the engine identifying the most appropriate communication in terms of offer, tone, copy, content and channel.
The instant outcome could be any of the following:
- Presenting personalised web content to an individual user - perhaps highlighting special offers or product bundles that the customer favours. This is a significant step ahead of conventional content management systems that may be able dynamically to amend copy and images by page but not by user.
- Directing an email to the customer about the product they're browsing, encouraging them to purchase and offering an incentive based on their past purchasing behaviour
- Instigating an outbound phone call or managing an inbound call that offers the customer help and advice on their purchase, based on the call centre operators' knowledge of past online and offline behaviour
- Sending a digitally printed and personalised mail pack including a brochure, based on the customer's ordering habits and preference for a 'mail order' approach
A real-life example for a travel company used offline data about a customer's future holiday and other product holdings to encourage customers, via DM and email, to log-in to a microsite which carried information specific to their holiday. As the customer engaged with the site this activity, combined with data held offline and decision analytics, directed email and web content selected from hundreds of options, delivering interactions that were both personal and pertinent and that ultimately resulted in cross-sales of other products.
A bank of interactions from differing web content to email executions is vital in truly realising the potential of real-time interactions, and over time brands can increase their variety of approaches to suit every possible outcome.
Synaptic Marketing
In a sense, the way this decision engine operates is as much a new marketing philosophy as it is a new piece of software. If a synapse is defined as “the point at which electrical signals move from one nerve cell to another”, this new philosophy is 'synaptic marketing' because of the way all available data is instantly linked together to derive a decision, just like in the human brain. And the decisions that are made and interactions assigned then become a 'memory' of how and why the customer has been communicated with, further improving subsequent real-time decisions and interactions.
The evolutionary leap to Web 3.0
Ambitious marketers are right now seeing the benefits of synaptic marketing, using the powerful combination of data, decisioning and cross-channel communications that embody this cutting edge approach. Smart marketers will also be looking ahead to identify and implement leading edge technology as a prelude to future advances in this concept, namely a fully realised 'Web 3.0'.
If Web 3.0 can be made to work, through even more advanced decision engines and a greater handle on the varied elements of data that need to be understood and collated, it will likely result in two types of value: direct value and indirect value.
- Direct value is generated from realising a relationship between two items that a human mind would not have connected; imagine receiving a new email from a potential customer, and having the application available to tell you that this is someone who you've met before (it was on your electronic calendar), and that they have friends that are near the same stage of making a decision to purchase products that you sell through your ecommerce site.
- Indirect value is derived from analysing the connections between objects in a data universe, typically using some kind of machine learning techniques. The data universe will not just be limited to a single organisation, but includes data from business partners and affiliates, as a result of companies creating new business models for themselves. For example, imagine collaborative filtering, but applied not just to the books or movie purchases you made via Amazon, but across your entire life. For an overly simple scenario, such a system might be able to suggest, based on the fact that you purchased swimming goggles, swimming trunks and read a lot of books on the Mediterranean, that an airfare ticket to the Mediterranean might be of interest to you. This is only possible in a world in which activity stream data is accessible across all of your applications, and is not locked into just one service.
It's important to note that the above examples, and indeed the entire thinking behind synaptic marketing, are based on the understanding that either the consumer's permission to use their data has been sought, or that the activity is based on non Personally Identifiable Information (PII). Adhering to the principles of permission and privacy are a vital prerequisite to any marketing activity, and even more pertinent when we reach this level of data-driven decision making.
Do androids dream of electric sheep?
These examples are still science fiction; and tellingly I don't believe that we yet have access to the kind of artificial decision engine needed to make these kinds of intelligent connections. But this approach - generating insight to direct relevant and meaningful interactions with customers - is already available, in the discussed synaptic marketing technique. The only difference here is the necessary (human) analytical element, which ensures that the scope of possible outcomes is comprehensive enough to meet the level of data input being made.
Synaptic marketing has the potential to revolutionise the way brands interact with their customers. Consumer apathy to traditional marketing, and the way digital channels and the internet specifically has brought instant communication to the fore, mean brands must change the way they approach marketing. And much like the proposed semantic web would aim to make sense of the vastness of the internet; synaptic marketing offers a means of humanising the interactions between brand and consumer, even if an artificial intelligence is at its heart.
Click here to view the article as it appeared in the magazine
Click here to download Adobe Reader for free
