Monthly Archives: April 2019

The Great Google Book Ad Experiment (Part 2)

Posted on by

(aka Can a Google Ad Sell Books)
Week 3

Google Ads

I decided to wait three weeks until I had some useful (?) data. If you remember from my last post the intention was to target people searching for Alternate History (or similar) using the Google search engine. The seven findings so far:

7 findings

  1. Despite specifically excluding search terms including bookshop or books, and only placing the ads before people search for Alternate History (or similar), Google decided to ignore my instructions because it needs to generate clicks not sales (see point 2). Specifically the phrase ‘books’ triggered the ad in 11,938 cases (nearly 1/3 of all impressions), and generated 15 clicks (out of a total of 108).
  2. Google tunes its placement algorithms based on the number of clicks, not sales. I concede that Google can’t actually identify sales (that’s my job) but its hardly a fair trial of my proposed strategy if Google does its own thing.
  3. Most of the ads placements were actually on Google partnered sites (97%), rather than in direct response to a search, once again preventing me from properly testing the effectiveness of my proposed strategy (see point 7(a) below).
  4. Despite only 3% of ads appeared on Google Search 13% of clicks came from ad placement on Google Search, indicating that such placements were 430% more effective than appearances on partnered sites.
  5. In addition, despite only 3% of ads appearing on Google Search they made up 43% of the cost.
  6. Advertising on a Friday is twice as expensive as it is on a Saturday, and three times what it is on a Sunday. That is $1.20 per click on a Friday, compared to $0.61 on a Saturday, and $0.47 on Sundays. OK – this one was an easy fix, I stopped advertising on Fridays and shifted that budget to the weekend, which should increase the number of click throughs by 50%.
  7. The Google Ads App has way more information than what’s available on the desktop – which is very weird. Information available only via the App includes:
    1. Where the App was viewed:
      1. Google Search – 3%
      1. Google Partner Sites – 97%
    1. What the App was viewed on:
      1. Smartphones – 39%
      1. Tablets – 47%
      1. Computers – 14%
    1. The Ad was clicked on:
      1. Google search – 13%
      1. Google partner sites – 87%

Next steps

  1. I have cancelled Friday’s advertising and transferred its budget to the weekend
  2. I have a phone conference scheduled in a couple of days with my Google Ads Campaign Specialist to discuss
  3. I see what happens over the next three weeks

The Great Google Book Ad Experiment

Posted on by

(aka Can a Google Ad Sell Books)

Week 1

Cover of Nightfall by Andrew J Harvey

I have previously blogged about the failure of online advertising to sell books. My most recent blog on this was in September 2015  and entitled: “Does online advertising work for books?” This blog was based on an IBPA survey, and our own experience in 2014 with a series of blog-ads that cost us $330. For this we got 86,000 views, which gave us slightly over 43 click-throughs, and which in turn resulted in two sales. Given other examples quoted by the IBPA it was clear that online advertising doesn’t work for books.

About two weeks ago, however, I received a phonecall from an individual working for Google to inform me that because I had registered Hague Publishing as a business on business.google.com  they would be providing technical and artistic assistance to me for three months to develop and tune a Google advertisement.

That generated a conversation that went for 1.5 hours as John and I tried to work out what products Hague Publishing produced (Science Fiction and Fantasy books), and what type of advertisement might actually result in us selling more books (probably none). However, during that conversation it occurred to me that online advertising of non-Fiction books such as self-help and DIY might result in some sales if the search terms were narrowly focussed on those areas the books were about. And then flowing on from that perhaps people searching for ‘alternate history’, which is a pretty niche genre, might be interested in an alternate history book. And so I agreed to advertise my most first book “Nightfall”, published by Zmok Books.

The first thing was to come up with the search terms that would be used. This is set by a combination of the ‘business’ and the services/products being sold. This was the first problem as describing the business as a ‘publisher’ gave terms that focussed on ‘printing’, not ‘books’. Changing the business to ‘bookseller’ created more useful categories, although even here terms such as ‘new books for 2016’ isn’t going to get your ad in front of the right people. Luckily Google lets you turn search terms ‘off’ which resulted in me turning off all 221. It was more useful however when I listed the services/products as ‘alternate history’. This gave me 20, very specific search terms.  

As the advertisement links direct to the paperback listed on the Hague Publishing site, and postage overseas is a real killer I set the advertisement to initially only be offered in Australia. A further restriction is that the advertisement will only run Friday-Sunday, hopefully putting itself before people who are in the mood to whip out their credit cards. In an effort to increase conversion rates, however, I also updated the listing to include links to eBook distributors, including Kobo, Google, Apple, and Amazon.

So will this advertisement actually appear before anyone, and if it does – will it result in any sales? I’ll keep you informed.