As more people buy voice-enabled devices, adoption of voice search is set to rise with some even predicting that it will replace text as the primary means of searching — like how smartphones overtook desktop.
Exploring the growth of voice search, an interesting search patent gives us a view into how it’s able to remove entities (i.e. people/things/items from consecutive searches), leaving brands without full knowledge of what people were searching for.
Search is more than an isolated task; people search for related topics at different times of the day, dipping in and out, but querying the same theme. On Google’s 20th anniversary, they said search is a journey, citing its ability to decipher and sync early searches in awareness through to a journey into research and purchase.
Voice search works on the same premise; we command and search answers. Where voice differs is timing; people are more inclined to seek answers over a shorter timeframe. To prevent people having to repeat themselves, in the true sense of a “search journey,” Google has patented a way to prevent anaphora. Anaphora — most commonly associated with politicians — is when people repeat the same words from their first sentence in their second. Politicians use anaphora to ensure their message is being delivered.
Keywords Without Entities
Applied to voice search, we can see how repeating words (anaphora) creates friction:
- Who is David Beckham?
- Who did David Beckham play for?
- What is David Beckham’s net-worth?
Three queries all related to the same entity, David Beckham, but ultimately different in intent.
Applying the notion of search journeys, Google can determine that your consecutive queries all relate to the same entity, thus enabling people to voice search without anaphora:
- Who is David Beckham?
- Who did he play for?
- What is his net-worth?
Notably, the second and third queries don’t include the entity, David Beckham.
People are able to search by voice more naturally, like how we’d speak to a person who’s knowledgeable about David Beckham.
Keywords are Essential Marketing Intelligence
This creates a challenge for brands: [who did he play for] is a longtail keyword considerably less insightful than [who did david beckham play for]. A football brand doing research into their audience wouldn’t know if searches were associated with Beckham, or in fact Ronaldo, Messi, Kane or any other player, for example. Again, per Google:
“… the first query can be ‘Who is Ben Franklin?’ The entity of the first query was identified as ‘Ben Franklin.’ The first query can be sent to a search engine. The search engine can return a list of resources referencing Ben Franklin. A second query can be received within a certain time frame asking, ‘What is his height?’ The pronoun ‘his’ was recognized by the search engine to be part of the second query.”
SEO search and PPC search both were built on keywords. Google Ads recently focused brands’ attention to audiences, encouraging them to bid on the person behind the search rather than the keyword.
The concept of keywords without entities isn’t new; it’s widely reported that search has already become more personal. Growth in [near me] and [best] searches emphasize this, showcasing that people want answers specific to them, but without having to enter [Edinburgh] or [London], as geographies from where they are searching.
Publishers Will Fill Brands’ Intelligence Needs
If the predictions are true, then we should expect that 30-50% of searches (Gartner and comScore) will be through voice. Paired with this new revelation, this could mean up to half of all keywords won’t contain an entity, thus eroding the insightfulness that keywords and their search volumes yield to brands and their marketing intelligence.
Searches provide brands with insight into seasonality for budget pacing, color nuances for designing products and market research, to name just a few examples.
Keywords have long been a staple of intelligence for brands, but their significance and level of detail has lessened over the years.
Positively, brands should be focusing on their audiences’ “needs” instead of vein metrics like traffic. Without keywords being as effective for audience insight, brands will need to find intelligence elsewhere.
We envisage further alignment between brands and publishers. Intelligence from publishers is currently mainly around shared audiences: What can you tell us about this audience? If voice search erodes keywords intelligence and privacy continues to tighten, to supplement for the loss of intelligence, publishers know a lot about their audiences through their first-party data which could yield great value to brands looking to attract audiences.
The future of voice search is conversational — a reflection of how people naturally ask questions. Brands need to reflect their expertise, authoritativeness and trustworthiness in content, a true reflection of an audience-first approach.
Originally published in MediaVillage.
Author: Michael Thomson, Senior Media Strategist, Digitas