How Keyword Length Performance Can Impact Smartphone Paid Search Strategy

As more and more advertisers expand their paid search programs onto mobile phones, it’s natural to wonder what keywords will be most successful. The differences in smartphone versus desktop/tablet SERPs, the clumsiness of users’ thumbs, and the large difference in CPCs between devices dictate that advertisers have to approach these spaces differently. Naturally, advertisers might be questioning if there are differences in how shorter queries perform versus longer queries. And with that data, what changes can we implement to improve performance in our accounts?

Keyword Character Length by Device

To help answer these questions, we investigated the performance of keyword character length for non-brand text ads on smartphones relative to desktop across a sample of RKG’s body of clients. Keywords were grouped into buckets of 1-5 characters, 6-10 characters, 11-15 characters, and so on up to 31+ characters. Starting by looking at click share by device, we found that the volume of searches was quite consistent across device types, with the biggest difference in click share (just 2%) coming from the 16-20 character bucket.

Keyword Length Click Share by Device

This, perhaps surprising, result indicates that users search habits don't differ between computers and smartphones as much as one might expect. This may be at least partially due to autosuggest completing users' queries for them, thereby making the searches on different device types more similar than would be the case if users had to type out their full queries. Regardless, click share by device should not affect smartphone strategies, because the differences are so small. Taking a look at sales per click (SPC), every character length bucket had a lower SPC on smartphones than on desktop, though keywords with character lengths between 6 and 15 showed the worst relative performance. Smartphone conversion rates were fairly similar across the character count buckets.

Smartphone vs Desktop by Keyword Length

Using Google's mobile modifiers, smartphone spend can still be efficient even with lower sales per click. Due to advertisers factoring in cross device conversions, new customer acquisition, and in store spillover, smartphone sales per click matches closely, although not perfectly with smartphone CPCs relative to desktops.

andy's cpc graph

Factoring in Exact Match

Viewing this data in isolation may not give advertisers the complete picture. If a keyword has a certain character length, but it is a broad or phrase match term, it may be matching to search queries that are significantly different in length than that keyword. Let’s look at similar data, but this time only for exact match terms. The click distribution by device for exact match keywords is very similar to the earlier click share chart, with the plurality of clicks coming from the 11-15 character bucket. However, exact match keywords do have different conversion rate and sales per click values relative to desktop by keyword character length than those for all match types.

Keyword Length Smartphone Performance vs Desktop - Exact

In this sample, very short exact match keywords between 1 and 5 characters had much higher sales per click relative to desktop terms than longer keywords, although this data may be anomalous. And, unlike the data including phrase and broad match keywords, smartphone sales per click relative to desktops improved in the 31+ character bucket. In aggregate, smartphone sales per click for exact match keywords relative to desktop was around 30%, compared to about 40% for all match types combined.

Leveraging Character Buckets on Smartphones

The relative differences between desktop and smartphone conversion rates and SPC did vary by match type, with broad and phrase match smartphone keywords having higher SPC and conversion rates than similar exact match terms, relative to desktop. This suggests that advertisers would benefit by segmenting their ad groups by match type, to take advantage of mobile bid modifiers at the ad group level and by being more aggressive with their bidding on broader match types on smartphones. Traffic distribution by keyword length is similar on both smartphones and desktops. Regardless of how we factor in match type, the differences in smartphone and desktop click share are small. This makes dividing keywords by theme and match type, rather than keyword length, the key to adgroup structure. By creating exact and broad/phrase ad groups, advertisers will have a clearer picture of the value of smartphone traffic relative to desktop, and can create more effective smartphone modifiers in their accounts.
Join the Discussion