The Reading Forecast

Sanoma Magazines implements SAS tools.

Tervuren, April 2011

SanomaIn the past you bought a magazine, and that was that. Today, you buy your favourite magazine often in combination with a CD, a good book or some other fun gadget. All very nice for the reader, but the broader range of options does make predicting demand a more complicated business. Ben Sprangers, ICT Business Solutions Manager at Sanoma Magazines, was confronted with this new reality and joined forces with Steven Raekelboom, external Master Data, Planning & Circulation Manager for Sanoma, to find a solution. They arrived at SAS’ forecasting, data management and optimisation tools, which they implemented in 2009, and are currently investigating the possibilities of multi-phase delivery.

Sanoma Magazines Belgium employs over five hundred people and has a turnover of EUR 223 million (2008). The organisation distributes around 25 mass-market magazines in Belgium and has a substantial market share in various sectors such as womens’, TV and home magazines. Some of its well-known titles are Humo (51% owned by Sanoma Magazines), Flair, Libelle, Story, Goedele, Feeling, Femmes d'Aujourd'hui, TeVe-Blad and Télémoustique.

The project focus was on reducing the volume of returns. Sanoma Magazines has to take back unsold magazines and also bears the costs of this flow of returns. Inevitably, the increasing diversification of the products also had to be taken into account. "We hardly ever sell just a magazine nowadays. Usually, a reader can choose between the magazine by itself or the magazine plus a giveaway, for example a DVD or another magazine. In fact, we often put out a magazine in three different variations nowadays, which only makes things more complicated. The previous distribution software wasn’t designed to handle this complexity", explains Steven Raekelboom.

"Most publishers have outsourced the entire control process, by which I mean the decisions taken on the distribution of magazines between the various sales outlets, to their distributor. We have an in-house Circulation team that tackles the distribution of magazines on the single-copy sales market. How many single copies of Humo are distributed and how many copies with a DVD? And what is the distribution between the 4,500 shop outlets that we supply with Dutch-language magazines? We want to keep this control process in-house because it gives us a degree of flexibility. It also keeps us more in touch with our products. Knowledge of the market undoubtedly plays a role in deciding the print run", says Ben Sprangers. "Plus the costs of returns are borne by the publisher. So it’s logical that we should want to control those flows ourselves," he adds.

The project’s first Key Performance Indicator (KPI) was reducing the volume of returns. To this first KPI was added a second, namely the ‘out-of-stock weighted’.
Ben Sprangers: "Suppose a magazine has sold out in 450 outlets: the out-of-stock is ten percent. We then look at those outlets’ share of total sales. If we’ve sold 150,000 copies on the market and those 450 outlets together sold 50,000 magazines, the out-of-stock weighted is thirty percent."
Steven Raekelboom: "It’s a constant balancing act between returns and out-of-stock. If you expect to sell 90,000 magazines and print 91,000, you’ll have low returns but a high out-of-stock. Turning it around, the same applies: putting out far too many magazines means high returns but few missed sales."

Quest for suitable tool

Sanoma Magazines included the two proposed KPIs in the project specifications. "How many percent fewer returns can we achieve, with the same out-of-stock level, by using a new tool? In other words, is it worth investing in a tool?", says Ben Sprangers.

Two IT suppliers were on the shortlist. In the end, the preference went to SAS.
B. Sprangers: "We were looking for a solution that is also strong on optimisation issues: given the demand in an outlet and given the volatility of demand, how do I distribute the magazines on the market in such a way as to minimise my expected losses? With its SAS Operations Research module, SAS is strong in this field. In addition, forecasting is a difficult business. You need to have a good understanding of how time series work before you can decide which method is most suitable for forecasting. The advantage of the SAS Forecast Server is that the system constantly tests a whole series of methods and selects the best. You don’t have to decide on the method yourself any more. So it’s kind of like industrialising the forecast."

In addition to SAS Operations Research and SAS Forecast Server, the publisher also purchased SAS Data Surveyor for SAP.
B. Sprangers: "At the moment we use Oracle, but from 2011 we will gradually change the entire organisation over to SAP. Our parent house Sanoma Finland, and Sanoma Netherlands, both use SAP and the intention is that we will all work on the same platform."

In the past, Sanoma Magazines didn’t use a real forecasting tool but a sort of database management system. "The solution no longer complied with the industry standard. And with the increasing complexity it wasn’t possible to keep doing everything in the old tool, because the tool required a lot of manual work. In addition, that tool had to be fed with data that we extracted ourselves from the databases, which was a fairly time-consuming process. A consequence of this was a limited history. We used four measurement points to determine demand. In our SAS application we include two years’ history, which leads to much better estimates", says Steven Raekelboom.

Go live

Sanoma Magazines’ management decided to sign up for SAS in December 2008 and went live around nine months later.
B. Sprangers: "We’d originally scheduled the ‘go live’ for June 2009 but the summer’s an atypical time in any case, which makes a go live difficult. In fact, the delay worked out OK for us. Our main concern was that the outlets shouldn’t notice the transition. So we didn’t announce in advance that we were going to be using a different tool for forecasting. We were looking out for more complaints to our service department, but that fear subsequently proved unfounded. In fact, the service department received fewer calls about control issues."

Under the bonnet, the SAS solution is more complex than the previous tool but according to Steven Raekelboom the users don’t notice it: "The users only see the specially developed user-friendly screens via a web interface. At the start I did have trouble ‘selling’ the automation of the optimisation issue. People don’t just believe that a tool can automate such issues. You have to win people over, give them the necessary training and gain their trust."

The Circulation team produces a forecast of the distribution between the various outlets, but the shop managers can overrule the forecast.
B. Sprangers: "A request like this is entered into the system as a temporary service, with a starting and ending date. Shop managers can have various reasons for placing a different order, for example due to local holidays of shops nearby or road works in the street. Of course, we carry out the usual analyses to see if the services the shop managers are requesting are acceptable. If a shop manager always requests a hundred magazines but, time after time, sells only forty, we will get in touch with him."

Visible results

As benefits of the forecasting tool, Ben Sprangers cites the flexibility of the solution and the reduced workload on the Circulation team. "Previously, the Circulation team devoted around sixty percent of its time to administration and forty percent to analysis. Today, that ratio is thirty/seventy. For me, it’s also very important that Sanoma Magazines is better armed for the future due to the implementation of a new forecasting tool. We are using two KPIs, but are they the right ones? SAS allows a better analysis. We now have an industry-standard solution in-house and we’re going to use it to the utmost to take our Circulation team to a different dimension."

"Typically, there’s an area of tension between the Circulation team and product management. Which is healthy. With the support of the forecasting tool, we don’t have to debate the figure any more, just the assumptions made. For example, we assume that it’ll be a top edition and we factor in that the magazine will be available in four different versions. Based on the history, given those assumptions, what will the demand be? You can debate the assumptions, but that’s easier than debating figures", says Steven Raekelboom. "In 2011 we also want to keep the history of the various control scenarios up to date. We produce several scenarios based on the different assumptions, but then, in consultation, choose a scenario that we market. The intention is to update the different scenarios to enable us to evaluate our decision afterwards. This allows you to systematise the business information and report on it more easily, and produces a healthier consensus model."


The project’s first Key Performance Indicator was reducing the volume of returns. Sanoma Magazines has to take back unsold magazines and also bears the costs of this flow of returns.

Sanoma Magazines didn’t give any specific figures on the savings made, but the Return on Investment (ROI) is good at any rate: the solution has paid for itself in under a year. S. Raekelboom: "The project was implemented in 2009. If you want to innovate at a time like that, taking account of the economic context, the criteria for an innovation pro­ject are very strict. We have met the criteria according to the measurements, although it’s a difficult comparison exercise because the first proof of concept dates from 2007. The market has changed enormously since then."

The future is multi-phase

Sanoma Magazines is now also investigating the possibility of multi-phase delivery to the sales outlets. Currently, it distributes several times a week to the sales outlets, but only once per edition. "On Tuesday it’s Humo in the crates, and on Thursday Libelle for example. But it’s not inconceivable that, instead of putting out the full print run of Humo straight away on Tuesday, we might put out part of it and distribute the rest only in a second phase. In the Thursday distribution round the crate will contain Libelle magazines but also, for example, a number of current-edition Humos. The volume stays the same, so the transport costs remain stable. This system has two major advantages in our view. First, it brings a greater degree of freedom because you can control on the basis of more information. That should reduce the return costs. Consider the coastal region for example. That’s a very volatile sales area because sales are strongly influenced by factors such as weather conditions. Second, the outlets don’t receive all of the magazines in one go, which should make for a better business in terms of organisation and presentation. Multi-phase delivery may also be a solution to the storage problems that shop managers face due to the increasing number of giveaways", says Ben Sprangers.
S. Raekelboom: "With our previous tool, a project like this would have been completely unfeasible because the number of control signals would double or even triple. But the idea of multi-phase delivery in itself has nothing to do with forecasting. Rather, it has to do with thinking about the supply chain. Forecasting is an instrument, not a goal."

However, it is not the intention to print the magazi­nes in several phases as well. "Phased printing of magazines is much too expensive. Instead of putting the full print run on the market, we will put out part of it and keep part of it back. We will distribute that part depending on differences in sales at the outlets", says Ben Sprangers.

"We also have an important changeover of ERP platform ahead of us. In that respect it’s vital to have good cooperation between business and IT. I believe the SAS project has shown that it’s possible to bring a project to a satisfactory conclusion provided you have good communication. By using the same IT system throughout the organisation, we also aim to participate more in 'knowledge sharing’. For example, it may be an option that the Dutch branch will use the same SAS tools in the future. For Business Intelligence (BI), we use Business Objects within the group. I think that, with the combination of SAP-SAS-Business Objects, we have good packages in-house that can safeguard our future", concludes Ben Sprangers.



About SAS
SAS is the leader in business analytics software and services, and the largest independent vendor in the business intelligence market. Through innovative solutions delivered within an integrated framework, SAS helps customers at more than 45,000 sites improve performance and deliver value by making better decisions faster. Since 1976 SAS has been giving customers around the world THE POWER TO KNOW®.

For more information contact:

SAS Institute nv
Gracy Poelman
Kasteel de Robiano
Hertenbergstraat 6
B - 3080 Tervuren
Tel.: +32 (0)2 766.07.52
Fax: +32 (0)2 766.07.77

SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and other countries. ® indicates USA registration. Other brand and product names are trademarks of their respective companies. Copyright 2011 SAS Institute Inc. All rights reserved.