Marketing in the age of platforms – thoughts from Turing Fest 2018

Last week saw an international crowd attend Turing Fest 2018 at the Edinburgh International Conference Centre. White Light’s Head of Digital Design, Matt, was in attendance and shares his takeaways from the marketing stream of the second day.

Lead image by Erika Stevenson

At this year’s Turing Fest there was a strong sense of a realignment in terms of how the internet works and a bedding in of previously cutting-edge technologies.

In particular, there was a fascinating thread that ran through a number of the presentations – namely those by Cindy Krum of Mobile Moxie, Jes Scholz of Ringier and Jono Alderson of Yoast.

Cindy discussed the way in which Google is moving from a crude understanding based around keywords to a more nuanced cataloguing of entities and concepts.

The metaphor she offered was that of a child who learns the concept of a table and a chair and how they are contextually linked. If that child learns another language, they only have to learn the new words, not take on the concept all over again as well. This facilitates search based on a better understanding of search intent but also allows Google to be more consistent in how it serves its non English-speaking audiences.

Tie this with the ability of AI to draw from existing datasets, take on corrective training from humans and seek patterns in a sophisticated and nuanced way, and things really get interesting. We’re looking at a world where search can easily connect disparate pieces of content with or without the most obvious context established by common keywords.

And Google has an army of humans performing these training processes all the time to improve its AIs toward specific ends. Every time you have to identify cars as part of a captcha, you’re making their self-driving cars a little better at recognizing other vehicles. Every time you see an answer box at the top the search results and click on the feedback, you’re training it to do a better job of giving one-hit answers to straightforward questions.

Now that we have we have a new model of understanding entities and the computational smarts to sort and connect it all, what’s the next step?

Disinter ... what?

The path is easiest to observe in Google’s growing practice of disintermediation. It’s an old term that harks from the banking industry but increasingly crops up now in relation to the web. Indeed, it cropped up during Rand Fishkin’s talk from Turingfest 2017, and it continues to be an issue. What it means in Google's case is that content and data from the open web will be presented within the SERP in order to answer questions more quickly and with a consistent presentation instead of simply directing the viewer through to the original source.

It’s arguably a boon for the user and, as Google gets better at this by both observing our behavior and taking on the training we give it in the form of feedback, it’ll lead to a much lower friction experience for the user.

But this is a headache for the site owner who is kept at arms length from the audience who nevertheless benefit from what they have published. The points of friction Google has sought to eliminate were the deal marketers made with their audience – we give you something you want and in return you take on some of our messages. In this search landscape, however, the organic results can see themselves relegated to a third class below paid ads and these featured snippets.

But while Google may be seen to have its thumb on the scales, search remains our gateway to the open web. The rise of voice search is set to alter this drastically though. By the end of 2018, the installed base for smart speakers is set to reach 100 million, and that means a lot of consumer voice searches. The landscape is somewhat different in the corporate realm but voice is still making inroads.

Now, while we’ve all become accustomed to the way in which marketing forms part of our online experience, there is no way to slip a keyword-driven advert into a conversational interaction with a smart assistant without giving the experience of talking to an absolute raving lunatic. And that’s something Google, for one, are going to take a hard pass on.

So does that mean the results in voice are something marketers can’t influence? Well, no. It just means, as Jono explained, that we need to return to the broader, classic view of marketing. 

“There is no way to slip a keyword-driven advert into a conversational interaction with a smart assistant without giving the experience of talking to an absolute raving lunatic”


A return to marketing fundamentals

Successful companies which have stood the test of time understand and invest in activities that build brand recognition and brand preference. How do we make someone ask Alexa about a ‘Dyson Cyclone V10’ and not about ‘handheld vacuum cleaners’?

As well as understanding entities, Google and others also have such a good understanding of us as to make us question if we’re being secretly eavesdropped on. This is a potential point of influence too: can we lay sufficient breadcrumbs in the form of online interactions with us and our content that would make Alexa or Google conclude that a customer would want to hear about our product before our competitors? 

This harks back to what Peter Houston of Media Matters said to us in White Light’s Got Issues about the death of what he calls the ‘cult of digital’, where marketers obsessively mined the seam of visibility offered by digital platforms, allowing the hacks to define their action rather than being focused on brand values.

We’ve had a long period where our websites could buy audience via platforms, either through paid promotion or by investing in smart SEO. But this is, in some ways, an anomaly in Google’s eyes. It doesn’t want to send people to websites, it wants to be the one to provide information – and it profits when users dwell in their platform.

Google has a stated aim to ‘organise the world’s information and make it universally accessible and useful.” As Cindy pointed out, the word ‘website’ does not appear in that phrase. So we can assume that wherever it can provide that information within it's platform, it's going to use all the tools at it's disposal to do so. 

Is all lost? Well, no – not if you have a product to sell or service which Google can’t easily layer itself on top of. If your model is to build what Jono calls ‘content castles’ – huge stores of knowledge which you monetise through advertising – then I’m sorry, but Google and Facebook are still here to eviscerate you.

Voice slow to conquer retail

On the retail front, there may be some breathing room for now. While consumers are being won over to the benefits of voice assistants, they appear to have been slow to incorporate them into their shopping habits. It’s a new behaviour and so we’d expect it to require a period of adjustment before people come to fully embrace the act of asking Alexa or Google to carry out transactions on their behalf. Perhaps there will need to be changes to the process to allay fears of inadvertent doll’s house purchases among other things.

Jes also talked about the direct applications of AI for marketing purposes – specifically the way in which, if we can cease to conflate automation with intelligence, we have the opportunity to not only save time but to actually improve on marketing KPIs.

There has always been a tension created by the ways in which technology, while opening new avenues, also makes old processes and sources of revenue redundant. As ever, businesses that operate online will find the only consistency is the fact that the online landscape continues to change. Adaptability and maintaining a meaningful connection with customers that engenders loyalty will offset the platforms’ moves to consolidate their ever-growing grip on our attention.

Elsewhere on the agenda we saw some great insight from Unbounce's Olie Gardner who described a much more thorough approach to personalisation where a few simple choices made by your viewer might lead to a significantly more tailored browsing experience. Perhaps the use-case examples or even testimonials would dynamically spring the correct business sector so as to be more persuasive, for example. 

He also underlined the importance of navigation. When viewers reach your content as opposed to your product pages, they may not fully understand what your company offers. The bog standard corporate website navigation structure rarely gives you this at a glance so they experimented with dedicated button-based navigation for their content pages which led to an uptick in visitors browsing through from content pages to product areas.

The keynote was given by Rand Fishkin formerly of Moz, and now busy building his new enterprise SparkToro. His focus was on launching products and, as well as discussing the nature of great teams – hint: it's not what any of the soundbite merchants would have you believe. 

He floated the idea that somewhere between the traditional high profile launch and the Minimum Viable Product (MVP) launch beloved of tech startups there is room for the Excellent Viable Product. This was described as a standard that works for smaller businesses who may have less to lose than the bigger players but who still have a reputation that would be jeopardised when loyal existing users were exposed to an MVP which may be seen as low quality. He also overlapped with Jono's presentation in that brand storytelling was a big part of how he likes to see new products launched – putting something people can connect with emotionally before talk of benefits and features.

All in all, it was a very worthwhile day and I'll certainly be looking forward to reviewing the presentations and booking in again for next year.