Google’s Instant Search: Some Considerations

Written by Sarah on . Posted in Technology

In the short time that has followed Google’s release of Instant search much has already been written about the possible ramifications for online marketing in general. Whilst it is certainly too early to assess the impact of giving users immediate feedback on their search queries, we can make an initial assessment of the considerations it raises for advertisers and affiliates.

It might seem obvious that as results now display instantaneously and change as the search terms are entered – without users having to hit enter after each search – it follows that users will refine their searches as they go, based on what they think of the instantaneously-served results.

However, it is not for certain that people know what it is they want to find before they start searching for it. Will users really continue refining their searches by using longer and longer search queries, or will they want to keep their options open and explore the web based on the results closest to a more limited string of terms?

For example, a user might know that they want to buy at 32 inch LCD TV, but beyond that they might not have a preference – or not have enough information to be able to decide – about the TV’s exact specification (brand, colour, serial number, etc). They might therefore decide to search on keywords related to ‘32 inch LCD TVs’, but not extend their Google search further than that. Any deeper comparisons may be done on the rest of the web, away from Google. If this is the case, we might wonder whether sites optimised for longer tail keywords are really in a better position following this change, as some have already argued.

Google’s attempt to predict with ever-greater accuracy what a user is looking for might be more effective on brand-led search terms than ones where generic terms lead. Compare the difference in results, for example, on ‘Comet sales’ against a generic term that is also related to a brand: ‘Low cost holidays’. The impact on sites that are dependent for their traffic on mis-spellings could be significant, but at the same time Google is good enough at predictive corrections to suggest results that have already anticipated the most common mis-spellings. Generic terms however are another matter. A company whose brand name incorporates a generic search term, such as ‘Low Cost Holidays’ or ‘Sports Shoes’ (or even part of a generic term), might find that users bypass the opportunity to click-through to their sites and opt to continue typing, refining their searches as Google automatically suggests results tailored to those generic terms and attempting more accurate and informed click-throughs. As such, Instant Search could potentially lead users further away from the advertiser quicker, and therefore it might be necessary for these advertisers to rank better for long-tail generic terms, or brand + generic long-tail terms (‘low cost holiday car rental’ or ‘low cost holiday travel insurance’, for instance).

Advertisers may find it necessary to look to affiliates to support them to a greater extent for long-tail keywords that follow common generic search queries. For example, at present Google returns results for Tesco, Curry’s and John Lewis on ‘LCD TV’; for Curry’s and Tesco on ‘LCD TV 32 inch’; but for none of these three on the longer tail ‘LCD TV 32 inch reviews’.

If we imagine a scenario in which users know roughly what they want – a TV, or travel insurance, for example – but not exactly what they want (or are not willing to search for exactly what they want) – a 32 inch Samsung LCD TV, for example – then searches are likely to stop somewhere ‘in the middle’: on a relatively short but nonetheless product- or service-specific search term. Users are unlikely to be satisfied with the results presented very early on in their Instant searches. A user beginning their search with the term ‘car’, for instance, does not necessarily want to shop at the first suggested results: Carphone Warehouse. Google’s Instant search results invite the user to extend their search far enough until Google has found an apt suggestion for their query. However, users may not wish to keep searching beyond what they are absolutely sure of looking for in case they miss potentially good offers or want to keep their options open.

We might compare this situation to an analogous one that has faced price comparison sites. As comparison sites have become more and more sophisticated, users are invited to refine their searches to an ever greater degree. It is no longer enough for a user to know they want an LCD TV, they are invited to go further and specify the brand, the colour, the dimensions and more in-depth specifications. As argued above however, it is not always clear that potential customers are yet aware of what they want to this extent, and even if they were, whether they would be willing to risk missing out on deals, offers, or similar products by going through a more in-depth search. One possible implication for affiliates might therefore be that the hard work of converting the sale would still take place on affiliates’ site, despite their route to this site from Google being slightly different, perhaps shorter.

There are however a few things that the appearance of Instant search makes more certain. One might be that the greater the use of Instant search the less important that below-the-fold results will be. This is likely in turn to privilege the paid results, which for most popular keywords with Instant search appear to be more prominent and more numerous than organic listings. Searches will be quicker, and so the impression times for each paid ad will be shorter, but the relative importance of the placement of these ads against the rest of the search results will be greater as the generic terms further down the page will get seen less. However, we might wonder whether people will still prefer to cast their eyes up and down the results page after each search, rather than only at the top few results with Instant search.

There is a final consideration to think about. If Google is expecting that users are more likely to continue searching for something until they find results which offer them what they are looking for, exposure to paid ads is going to be far greater, even if they are not clicked on precisely because the paid ads change as the queries change. For example, suppose a user wants to search for ‘mobile phone exchange’ with the intention of getting cashback on their old handset. The keyword ‘mobile’ might return paid ads for Carphone Warehouse, as might the extension to ‘mobile phone’. Finally, if the full search term, ‘mobile phone exchange’, also shows a paid ad for Carphone Warehouse the user will have been exposed to Carphone Warehouse ads at every stage of their search – essentially, for three different search terms. Surely this would increase the chance of them clicking the Carphone Warehouse ad, even if its position was lower than a competitor’s – such as Mobile Phone Xchange’s – given the likelihood that the user now has the impression that Carphone Warehouse is the best possible result for their mobile-related query? Carphone Warehouse could be embedded in the user’s mind as synonymous with mobiles.

Search habits and attitudes vary from person to person, as does the way we interact with our computers when we search. Think, for example, of the difference in user experience for someone who touch-types and can look at a page of results at the same time as they search, compared to someone who looks from the keyboard to the screen alternately. The introduction of Instant search by Google might be seen as a part of its wider strategy to bridge the gap between what the user looks for and what they find, and therefore an attempt to keep users on Google and diminish the time spent hopping around the rest of the web. We will have to wait and see however to what extent users adopt the new service or reject it, and how affiliates and advertisers adapt to this new way of conducting search.

Owen Hewitson

Client Strategist

Tags: , , , , , , ,

Trackback from your site.

Comments (2)

Leave a comment