Attire Accessories - Feb/Mar 2019 (Issue 74)

DAVID FAIRHURST 113 Speech therapy David Fairhurst gives us a crash course in voice search engine optimisation Back in the 1980s, I worked for an engineering company that made precision parts for trains, helicopters and all sorts of high-tech, extremely secretive military projects. All this technical wizardry was created by a CNC machining centre that was programmed using a punched paper tape machine, allied with a green screen computer system with less processing power than the average modern watch. How things have changed. Nowadays, smartphones have more processing power than the supercomputers used by NASA to land men on the moon, and the devices we use are only getting smarter every year. These are fed by enormous amounts of freely available data, AI systems and hugely powerful multi-core processors. Now we all face a life dominated by AI-driven voice-activated search, things can never go back to what they were. If you want people to find your website using voice command search technologies, read on. OPTIMISING FOR VOICE SEARCH Before the development of voice-activated search, the keyphrases we all put into search engines tended to be driven by the technologies of the time – keyboard input in most cases. This means keyphrases tended to get truncated as humans adapted to the way search engines operated. Instead of typing “where is the best place to have breakfast in Caernarfon,” we would type “breakfast Caernarfon,” purely because it worked just as well but maybe because we all secretly hate to type! For the large percentage of UK searchers using Google as their main engine, the inclusion of the Universal Search update a few years ago meant that Google would now understand the longer, more conversational search queries we use, leading to longer search queries as the norm. Now that we have voice search, normal speech is recognised to a high degree and used by search engines to return appropriate results that would have returned large amounts of junk prior to the introduction of AI systems and said voice search. Knowing that the AI systems powering voice-activated search systems understand human speech and can even adapt to regionality tells us how to optimise for voice-activated search. Where previously with non-voice search a small range of keyphrases would be chosen for each optimised page within a website, queries for