The power of conversational search

Hey, Siri. Are mar­keters and their busi­ness­es invest­ing enough time, mon­ey and effort in improv­ing their con­ver­sa­tion­al search rank­ings? 

Giv­en that Juniper Research fore­casts that con­sumers will be using voice assis­tants on more than 8.4 bil­lion devices by 2024, while Mar­ket­sand­Mar­kets pre­dicts that the glob­al con­ver­sa­tion­al AI mar­ket will grow from £3.4bn in 2020 to £9.8bn in 2025, there’s a strong case that they should be doing more.

Firms that have already invest­ed in achiev­ing high­er voice search rank­ings are ben­e­fit­ing from it. For instance, trans­la­tion ser­vice Lion­bridge start­ed opti­mis­ing for voice search in July 2019. Twelve months lat­er, it had almost 47 times more ‘fea­tured snip­pets’ – the short text at the top of a page of Google search results. This improve­ment aid­ed a 127% year-on-year increase in traf­fic to the firm’s web­site.

“Busi­ness­es should view opti­mis­ing for con­ver­sa­tion­al search as manda­to­ry,” argues Olga Andrienko, head of glob­al mar­ket­ing at Sem­rush, which man­ages Lionbridge’s mar­ket­ing ana­lyt­ics. “We have already seen the dra­mat­ic effect it can have on busi­ness­es’ search results. As voice assis­tants are fur­ther inte­grat­ed into people’s every­day lives, this influ­ence will grow.”

Nick McQuire, chief of enter­prise research at tech con­sul­tan­cy CCS Insight, agrees. “As one of AI’s most impor­tant areas of devel­op­ment, con­ver­sa­tion­al search is pro­gress­ing rapid­ly,” he says.

Bringing AI to the masses

Con­ver­sa­tion­al search devel­op­ment has accel­er­at­ed “because it sits at the cen­tre of two impor­tant trends”, McQuire sug­gests. First is the improve­ment in AI speech tech­nol­o­gy. Sec­ond is the need to improve infor­ma­tion search in busi­ness­es. “This area”, he says, “is often list­ed as ‘bro­ken’ by cus­tomers owing to infor­ma­tion silos, espe­cial­ly across sev­er­al data stores, doc­u­ments and appli­ca­tions.”

McQuire con­tin­ues: “The fact that all the big tech firms have start­ed to tack­le this area with prod­ucts and tools demon­strates the scale of the cus­tomer need.”

Con­ver­sa­tion­al search is a more com­plex mat­ter than peo­ple sim­ply using their voic­es instead of key­boards, though. For instance, on what smart device is the search being con­duct­ed – a screen-less Apple Home­Pod, a Google Nest Hub Max or an in-car Ama­zon Alexa?

While com­pa­nies can pay to appear on the first page of a con­ven­tion­al typed search on a com­put­er screen, mas­ter­ing con­ver­sa­tion­al search is not so straight­for­ward. For one thing, mar­keters don’t always have the same real estate for adver­tis­ing. 

If a device being used for voice search has no screen, how like­ly is it that a busi­ness will pull traf­fic to its web­site if it ranks out­side the top three search results? Equal­ly, if the device has a screen, a more visu­al response is pre­sum­ably bet­ter. Notably, the aver­age answer length for the three most pop­u­lar voice assis­tants – Siri, Alexa and Google Assis­tant – is 23 words, accord­ing to Sem­rush.

Convenient and contextualised answers

Seman­tics aside, the cru­cial point for mar­keters is that con­sumers want their prob­lems solved quick­ly. So says Kashif Naqsh­ban­di, CMO at IT recruiter Tenth Rev­o­lu­tion Group, who adds: “Most of the time they are not just look­ing for a ‘what’, but also a ‘how’ and ‘why’. This is why ques­tion-dri­ven con­ver­sa­tion­al search has spiked.”

Naqsh­ban­di explains that few peo­ple are now search­ing online for, say, a “moun­tain bike” and then click­ing through web pages of infor­ma­tion. Instead, they are ask­ing spe­cif­ic ques­tions – for instance, “what sort of moun­tain bike should I buy?” – to seek a con­tex­tu­alised answer.

“Con­sumers are more accus­tomed to these fast, per­son­alised inter­ac­tions that help them cut through the dig­i­tal noise and find a solu­tion,” he says. “Mar­keters have to evolve to deliv­er inter­ac­tive, dia­logue-based expe­ri­ences to stay com­pet­i­tive.”

Euan Matthews, direc­tor of AI and inno­va­tion at Con­tactEngine, a devel­op­er of con­ver­sa­tion­al AI sys­tems, agrees that peo­ple ulti­mate­ly want to save time wher­ev­er pos­si­ble and are hap­py to pay for con­ve­nience. 

“Con­sumers embrace new tech­nol­o­gy if it makes life sim­pler for them. Oth­er­wise, what’s the point?” he says, attribut­ing Amazon’s con­tin­u­ing ascen­dan­cy to the time it saves cus­tomers. But he adds that a “big pit­fall” for mar­keters is to focus only on the search ele­ment.

To illus­trate his point, Matthews says that some devices, when asked about the best local Indi­an restau­rant, say, will now offer a fol­low-up option of book­ing a table through Google, and – if the user accepts – they will add this book­ing to the dig­i­tal cal­en­dar. 

“This time sav­ing is not because of the con­ver­sa­tion­al search,” he says. “It’s more because that search has been seam­less­ly mar­ried with the abil­i­ty to exe­cute a trans­ac­tion on your behalf. Mar­keters must con­sid­er how to mar­ry con­ver­sa­tion­al search and the trans­ac­tion­al capa­bil­i­ty of the voice assis­tant, because this is what saves con­sumers time and makes it more like­ly that they’ll progress down the sales fun­nel.”

The direc­tion of trav­el is clear, so mar­keters must alter their course accord­ing­ly. “We will see more advances in the abil­i­ty for con­ver­sa­tion­al search to end in a trans­ac­tion – and this will dri­ve uptake,” Matthews pre­dicts. 

Will it ever replace keyed search? “I doubt it,” he says. “Some search­es are best not voiced aloud.”