THE TOOL CHEST

What We Learn with BookScan Data

by Christopher Robbins

“God is in the details,” said the architect Ludwig Mies van der Rohe. More than ever before, the publishing community has the ability to analyze and pay attention to the details, to get granular with data, to spot trends, identify anomalies, and either leverage the positives or make course corrections for the negatives.

While publishing is a lot about art, I believe a good combination of gut feel and statistical analysis helps make for better decisions. At Gibbs Smith, BookScan is an essential tool for industry analysis, and I want to share information about how we look at the data and how we use them.

I asked Brianna Buckley, my liaison at BookScan, to give me her own pitch. She said:

“Nielsen BookScan is the world’s largest continuous book-sales tracking service operating in the United States, Australia, South Africa, the United Kingdom, Ireland, Italy, Spain, Denmark, New Zealand, and India. We collect total transaction data at the point of sale directly from the tills and dispatch systems of all major book retailers. This ensures that detailed and highly accurate sales information on which books are selling is available to the book trade. The United States Consumer Market panel covers approximately 75 percent of physical retail sales and continues to grow. In a typical week, over 500,000 different ISBNs are tracked, selling approximately 13 million units in total, and sales information is available across the total market or by retail channel. Data are available to subscribing publishers and contributing retailers on a weekly basis, just four days after the period end, via a proprietary Web site. Sales reports of units sold by week, year-to-date, and historical are available, along with sales searches by title and author. Bestseller charts by category, by region, and by DMA can be accessed.”

And she ended her pitch with a reminder: “IBPA members with annual revenue under $3.5 million are eligible to subscribe to this service at a significant saving of over 80 percent.”

While Brianna suggests that BookScan represents approximately 75 percent of the market, we did our own analysis and came away believing that, on average, BookScan represents about 45 to 55 percent of our business. (When Borders declared bankruptcy, that percentage changed.)

This difference occurs because Gibbs Smith has a significant specialty business, and a big percentage of our sales are not tracked through BookScan, which collects figures mostly from retail outlets that specialize in selling books and/or sell large quantities of books.

I recommend that you figure out how much of your business BookScan data capture before you feel giddy or queasy about its sell-through figures.

Significant Snapshots

Each week Gibbs Smith analyzes its BookScan data, looking first at four figures for our own company and comparing them to figures for the industry.

We check:

• week-over-week sales

• weekly comp sales to the prior year

• calendar year changes

• fiscal year changes

This snapshot approach allows us to see whether our retail sales are trending up during specific promotions or seasons; whether they are improving from the previous year; and how we are faring for the calendar year and fiscal year.

For obvious reasons, we benchmark these figures against our own goals and against the industry.

Then we graph overall sales data and weight them according to our own genres, because we don’t want to compare our own data with data that include sales in genres we do not publish.

Here’s the graph for the week just past as I write:

It’s always fascinating to see how closely our own sales parallel consumer activity in the BookScan reporting accounts. As our accounts fare, so BookScan respondents fare.

Of course, we are always aiming to outperform the industry. Monitoring the data helps us identify promotional activities that may have positively influenced our sales, which we can then repeat.

Pointers on Progress

We also monitor our retail sales according to what we believe we need to sell through to meet our annual sales goals. By extrapolating what our vendor-inventory turnover objectives are and how fast inventory needs to move through vendors’ cash registers to get them to reorder from us, we can produce a graph such as the one below.

Like the pool markers for an Olympic swimming event, this graph shows us where we are in relation to our objective. In the period represented, the large spikes reflect last year’s holiday sales and our own holiday unit goals, and we are underperforming to goal but slightly ahead of last year.

As Amazon Ascends

We also do an analysis to compare data about our major customers. When Borders was alive, we used BookScan to create gap analyses that helped us influence distribution. We still find it valuable to compare Amazon to Barnes & Noble.

What’s interesting this year is that, for Gibbs Smith, Barnes & Noble is dramatically losing market share to Amazon. We sold more books through B&N than through Amazon in only three weeks this year, although B&N was consistently ahead of Amazon by 20 percent in 2010.

What is going on? The data indicate that Borders’ customers did not migrate to B&N as hoped, which is a concern. The data also indicate that physical book sales are migrating from B&N to Amazon, which is helpful to know. We can see that we want to adapt our business to be more diversified in terms of distribution and to adapt to a changing marketplace by deploying more content digitally to take advantage of the fact that B&N is devoting more space to the NOOK.

What the data show is critical when we’re working with B&N’s buyers and considering co-op placement, not to mention when we’re focusing on leverage with Amazon. As we all know, Amazon makes many of its publisher marketing decisions based on the terms it receives from a publisher. When we see our Amazon receivables going up disproportionately and our customers buying more books through that account, we know we should be thinking about broadening distribution so that business decisions are not overly leveraged by Amazon’s influence.

While dealing with many averages in this analysis, we pay particular attention to price-point data. Our analysis of our cash register sales proved what we expected to see—that the average list price per sale went down for almost all our partners except Amazon, which discounts so heavily that more expensive books are more affordable to a broader market.

Data showing that most of the market is going to be Amazon indicate that we have an incentive to build Amazon’s discount into the price of the book. These data also allow us to explore our pricing strategy—are we leaving money on the table, or is the market still price sensitive? Current analysis suggests that if a book is unique to a niche market, demand is unaffected by price. However, if the book is more like a commodity, demand becomes very elastic in relation to price.

We also use BookScan data to identify positive sales anomalies. Each week we identify any title or titles whose sales increased by more than 50 units and discuss what we can do to repeat that pattern. BookScan allows you to dig into the region and city for those sales, so the effects of regional publicity and events become more measurable.

Clues for Success in a Stagnant Time

Our final, and perhaps most meaningful, measurement is of the entire industry’s unit sales growth on a monthly basis. As I said in “Action Items Fueled by BookStats” (November), these data provide conclusive evidence that the physical-book industry as a whole is on a week-over-week downward trend, with some genres more affected than others. The slight uptick from July through September is due to Borders’ bankruptcy.

It’s critical to understand that the only meaningful growth we’ve had in the print book market for four years was a direct result of liquidation. So the question we continually ask ourselves—and the question each publisher must answer for itself—then becomes, how can we dramatically outperform the industry?

Mies van der Rohe also famously said, “Less is more,” but more data are helpful when they let you see what is happening to your own business in relation to the entire industry. The information allows for greater flexibility in making more competitive decisions. And in this market, we need to be far more competitive than ever before to succeed.

Christopher Robbins, a member of the IBPA board, describes himself as a husband and the father of nine children who moonlights as the CEO of Gibbs Smith.