It is clear that now anyone can create and distribute their book to a global audience of millions in a matter of seconds, that successful marketing and promotion is the single activity that will define institutional publishing in the 21st century.
However, book marketing is broken. It is not evolving at the rapid speed of other parts of the industry, and many campaigns look identical to those 10 or more years ago, relying on PR and big-budget poster campaigns.
Why do campaigns stick so rigidly to this formula?
There is not – nor has there ever been – any empirical understanding of which combination of marketing strategy and tactics works. Currently, publishers might spend between 5-10% of their revenues on merely hopeful marketing, with little idea of what works – and what doesn’t.
This is a huge problem that is only going to get worse as the retail and promotional landscape deteriorates.
We firmly believe the answer to this problem lies in “big data” – that publishing can dramatically improve its return on marketing investment through the methodical analysis of campaigns, in combination with more agile marketing techniques.
If you consider the decision-making approach of new publishing entrants such as Amazon, Apple and Google, it is clear that publishing houses still make their decisions based largely on gut feel rather than on data.
However, it’s not so easy: the data that is available to publishers is becoming increasingly irrelevant. Weekly Nielsen sales data gives no insight into ebook sales; does not break down sales by day or by hour, or by location; and such broadness makes it impossible to separate the impact of one activity from another in the course of a week.
At the same time, Amazon rankings have become a proxy for performance, and we heard many stories of people obsessively refreshing their page on Amazon for minute variations in sales rank.
So, after researching how poorly publishers were quantifying the new world, and believing that a data-driven understanding of consumer behaviour is fundamental to the future of the industry, we decided to stop making apps, and to focus instead on building a market intelligence service for books that we have called Bookseer.
Bookseer captures the real-time data “exhaust” of the web, combines it with promotional information provided by the publisher, and builds a picture of the variations in performance of thousands of titles.
Bookseer collects as much information about these books as possible. It captures price on Amazon, hourly sales rank, print and ebook sales data (uploaded by the publisher), what is being said about a book or author on Twitter, Facebook or in the media, and on the wider web, which marketing is being engaged with, the makeup of bestseller charts and so on.
We collect this data in a variety of ways, but we do it all in real-time. This means that so long as we can identify significant events – media coverage, a tweet, a price change, a review – at a specific moment in time, we can measure the effect of that event on sales.
Let me show you what I mean with a couple of quick examples.
The first is a big-budget marketing campaign from the UK.
We tracked all 7 titles written by a well-known author, in digital and print format, in advance of a nationwide poster campaign (with a social media component) that promoted 3 titles.
The remaining 4 titles were tracked as a “control” to measure baseline performance on the non-promoted titles. We also tracked mentions of the author name, and the social media keyword promoted on the posters – all of which probably cost in the region of £75- £100,000.
As well as tracking Amazon sales rank, price and social media, we included Nielsen sales so we could see how closely Amazon sales rank and actual sales are correlated. The blue lines at the top represent the hourly sales rank (noting it’s incrementally harder to get higher up the chart), the blue bars are the actual sales through Nielsen, and the red lines are the Kindle sales rank.
The first circle here is when the campaign started, and ended here, at the second. As you can see, the run of the poster campaign made very little impact on sales, although if one were generous, one could attribute a slight increase in actual Nielsen sales (of around 150 copies), and a slight stemming of the downward trend of sales rank during the campaign.
Here we can see the social media metrics.
The blue line is the number of Facebook “fans” for the author, which is steadily climbing, but which remains unaffected by the campaign, again shown between point 1 here and point 2 here. There is a spike in the instances of the Twitter hashtag promoted on the posters, although the actual number of tweets generated was 37.
I’ll let you make up your own mind as to whether this was worth the spend.
My second case study also highlights the risks of big-budget marketing. This data was collected for a chick-lit title in the UK, with a heavily digital launch spend.
Ads were placed on YouTube, and a high-traffic community website was targeted with a strong demographic crossover with the book. All advertising space on the website was bought, pointing to the Amazon product page.
Furthermore an author livechat and two direct emails to site subscribers – over 350,000 of them – were secured. Estimates are that this was a £30-50,000 spend.
So, here on Bookseer we can see when:
1 – advertising goes live on site
2 – Live chat takes place
4 – Two emails were sent to groups of 175,000 subscribers. (I’m going to zoom in on the data on an hourly basis so we can see exactly when the emails were sent here, and here
Returning to the less detailed view, more obviously, we can see the impact of
3 Which is when the author appeared on BBC breakfast and was interviewed in the Evening Standard. Going back to the detail view we can see that BBC had much more effect than the Standard, but in combination they both delivered great returns.
Again, make up your own minds – the question we want people to ask is, “With this data, would I spend my money in the same way again? How would my campaign differ?”
These are just two examples. We have other case studies that we think overwhelmingly demonstrate the need for a new approach to book marketing.
This approach is one that (obviously) captures and analyses data, but more fundamentally is simply more agile.
Rather than believing that expensive media space booked months in advance is the best marketing tool, the new marketing spreads its bets wisely. It monitors and hones multiple activities on a daily basis to find the right mix of PR, social media advertising, search engine advertising or direct mail. It looks at broad ranges of data in real time to see which activities work, and tweaks and refines the activities and propositions through split A/B testing and audience segmentation. And it adopts its learnings into future campaigns as it goes.
By being hands-on in this way (as opposed to just throwing money at the problem) publishers can monitor and refine a failing campaign into one that delivers much more ROI.
Bookseer is designed to demonstrate the impact of any marketing approach, and to quantify it against the metrics that matter in publishing: more sales.
But simply measuring isn’t enough. This new marketing approach undoubtedly calls for significant training and investment in skills, skills that publishing does not currently have, but which are fundamental to delivering the consumer-centric vision of the major houses around the world.
Let me finish by articulating our vision.
In the short term, a data-centric approach allows publishers to be much more reactive to an individual campaign and to tailor their activity to the bits that are working, and stop those that aren’t. They can compare the impact of PR against marketing spend, and measure the outcome of a price change. They can see what worked on a competitor’s title, and learn from it for theirs.
But, more excitingly, in the medium to long term one can build up an incredibly large corpus of data that can then be mined algorithmically to run predictive scenarios for books you will publish, using data from those that have already been published.
We believe that this data can inform future publishing strategies: what is the optimum price, format, time of year to publish? Who will the most successful retailer be? Who are the top 5 journalists who have the best track record of driving sales in this genre? Who are the most influential social media people to reach out to? What will the sales be, and therefore, what is the right advance to pay?
These are the questions fundamental to bringing books to market, and it’s time that data played it part in answering them.
Thanks very much for your time and attention, and for allowing me to do this on video.
If you have any questions I’ll be online now, and I’d be delighted to answer questions via twitter either to @gunzalis, or with the #bookseer hashtag.
Bookseer is currently in closed beta. Please contact Peter Collingridge for more details.