Click: the future of digital

In this regard, we must have a sense of the direction of things without being committed to a specific path – the ability to deviate from a stated direction is as important as having a direction in the first place. With this proviso in place, however, the following are some of the strong opinions that I currently (if loosely) hold regarding the future of search.

How digital will look in five years’ time

A more even market share will require more thorough cross-platform strategies

At present, Google’s market share seems unassailable, but companies have held such positions before. In five years’ time, we believe the search industry will be required to place more effort into strategies for a more fragmented and specialised search landscape.

Bing, for example, is making great leaps in search and their organic and paid offering is almost on par with that of Google. If Bing and one or two of the smaller search engines, for example, Yandex or Baidu – the largest non-US search engines, are able to specialise for a single type of search or service and capitalise on a USP, the search landscape could look a lot different. Bing’s visual search is, so far, superior to Google’s, for example.

In this different landscape, it will be necessary for digital marketing to move away from what has been an almost exclusively Google driven development path and look to build unique strategies for different query types and search engines.

The drive to answer questions in search engine results pages (SERPs) will see traffic fall but relevance rise

A common controversy in search at the moment is the push (specifically by Google, but by all search engines more generally) to answer queries directly in the SERP using various rich result types. In five years’ time, we are likely to be no closer to a truly equitable way in which Google redistributes its gains to the sites it takes the content from for these rich results, but digital marketing will have to develop bespoke strategies to create content for these rich results that develop brands in informational queries while blocking such rich results from pages of financial importance to the brand.

While traffic to brands may drop, the brands that are able to successfully compete for the remaining share will see the relevance of this traffic to their brand increase even as they see a reduction in volume, as they implement better content targeting strategies to develop genuine relationships with returning consumers and to build relationships with new consumers.

The importance of schema mark-up will no longer be deniable

The internet is the largest library the world has ever produced, but the Dewey Decimal System is insufficient for a library that is expanding exponentially. As such, the effort to create a searchable, sortable distributed database has required the development of a new language. The open source schema project represents the best way we currently have to properly organise information on the web. With new vocabularies in development all the time, we are approaching a point at which we can assist the organisation and categorisation of the web and help search engines return the best results at all times.

The importance of schema to machine learning models or category and entity detection as well as to blossoming voice activated searches will render schema mark-up and other information architecture strategies vital for brands looking to succeed online.

Smart speakers will reach 2/3rds adoption in the UK, meaning brands will need to be present

The Smart speaker reached 50% market saturation faster than any other device including the smart phone in the US. In the UK, progress has been slower, but smart speakers have still found themselves in 20% of all UK households. As their usefulness increases and as people begin to see them in their friend’s homes and begin to have conversations about them, adoption is likely to rise and do so relatively quickly – reaching 66% of the UK within five years.

As a result of this, brands are going to need to ensure they’re present and discoverable on these new devices.

How digital will look in ten years’ time

Federated learning will allow for personalised UX without direct data collection

The computing power of the current generation of phone is more than capable of supporting the federated learning method – wherein a user’s device is given the learning model, and then all of the learning (but no actual user information) is uploaded and used to educate the next iteration of the global model – which is downloaded to the user’s phone for the process to begin again.

With the increase in computing power combined with a number of open source ML modelling options (including Google’s own TensorFlow), brands will be able to create or license their own ML that will interact with such a federated learning program to offer personalised user experiences to consumers without requiring their data - only their input and interaction - in effect becoming more 'predictive' than 'personalised'.

Voice search will have reached maturity

Voice search is missing one or two key items of technology to be truly a part of our lives – but they are in development. One of the key technologies – a wearable subvocaliser – is already in prototype stage at Microsoft, IBM and MIT. This will allow us to communicate with voice activated search silently and often – overcoming many of the barriers inhibiting its widespread adoption. This will mean that digital marketing will need to move to a world of voice first just as it did to mobile search – but with the understanding that many of these searches will result in 2-3 answers or a shift to a screen, so that every query could end up being multi-device in nature.

ML will humanise our digital assistants, causing a level of change similar to the smartphone

Our current digital assistants are the clunky, black and white television versions of their future selves. Their functionality is minimal, their ability to track search intent through multiple follow-up queries is poor and their ability to mould themselves to their user is non-existent. All of these issues that render them only a step above novelty as things stand are likely to be resolved by various ML models that are already in use elsewhere, but which are too resource intensive for the current generation of digital assistants.

However, the developers at Google have created a way to use cloud computing to allow a small executable file on a smartphone to pull the much larger and smarter Google Home Assistant on to their next generation smartphone. Over the coming decade, these assistants will become better equipped to function as actual assistants and learn to adapt to their user without central data collection and are likely to be the next major technological revolution, playing a part in our health and wellbeing, our day-to-day routine and demanding that our content is laser focused on our chosen demographic to ensure it reaches the consumer past these digital gatekeepers.

Subject to change

Obviously, as stated in the introduction, the trends from which I’m inferring these developments could - and almost certainly will - change. This is part of the reason Click Consult runs its annual Benchmark Search Conference. Every year we look to gather together some of the people we think are able to best elucidate the opaque future of our industry for us and the delegates that attend.

It’s for this reason we were delighted that Richard Wallis (one of the founder members of schema.org) will be joining us this year and that Vodafone’s Nick Wilsdon – one of the country’s most knowledgeable practitioners of voice search techniques – returns. Combined with a host of exciting speakers, we’ll be hoping to either change our minds or continue in our current direction for the next year with a little more confidence.

If you’d like to join us, you can apply for a ticket here.