In an exclusive Q&A interview, TRBusiness interviews Generation Research’s owner Fredrik Lindh on how the company is approaching its DF&TR data collection, analysis and forecasting methodology today and what the company’s new partnerships bring in terms of new resources. Doug Newhouse reports.

 

 

WHAT DEDICATED DATA AND RESEARCH SALES PRODUCTS ARE YOU CURRENTLY PROVIDING AGAINST THOSE THAT YOU HAVE PRODUCED FOR BOTH SUPPLIERS AND RETAILERS OVER THE LAST THREE DECADES?

Fredrik Lindh, Generation Research Owner: Currently we are providing sell-out data in volumes and value for Perfume and Cosmetics, Wine & Spirits, Tobacco and Confectionery down to SKU levels.

 

We have other category products such as TRPR for beauty and TRIQ for tobacco and confectionery, where we produce the data with the help of all participants (brands) who provide their own individual company data and which we consolidate without revealing the individual company information.

 

 

 

 

IN EFFECT, YOU ARE ESSENTIALLY RELYING ON THE BRANDS TO SUPPLY YOU WITH THE ACCURATE DATA?

Partly yes, because some airports and retailers who share detailed data with us do not accept that we go down to country or airport level. This is inherent to the concession business we are in.

 

 

 

SO IN REALITY, DO YOU ACTUALLY HAVE ENOUGH RETAILERS INVOLVED IN THIS SYSTEM OF SHARING THEIR SELL-OUT DATA?

Through decades of negotiations with retailers and airports we now have 40% of global retailer data overall with sell-out data on different levels. This is a good base to indicate where the business is heading, but it is not enough.

 

Our mathematicians/data analysts estimate that we should have at least 60% to be able to fully rely on our retailer-based panel. The panels are up and running and we are working on improving the participation rate.

 

 

 

 

CONSIDERING ALL THIS, EXACTLY HOW DO YOU ESTIMATE THE TOTAL DATA CURRENTLY?

This is a unique expertise that we have developed over almost 30 years being dedicated to doing just this: estimating missing sales data for the TR industry. It relies on three primary processes. The first is other sources of data, and we have many.

 

To name a few, we receive data from nearly 80% of the suppliers (brands), from individual airports, airlines, ferries etc.

 

We scan the press, all public information such as government statistics, company annual reports and airport product assortment checks to name a few.

 

The second is a very complex mathematical model, which we have developed in-house and have fine-tuned over the years, and the third is a team of category specialists who are dedicated category experts in the travel retail business.

 

 

 

 

OK, BUT MODELS USED TO ESTIMATE OR ROUND OUT MISSING DATA ARE WELL KNOWN AND DOCUMENTED AND ALL DOMESTIC PANELS DO THIS. WHY AREN’T YOU USING ONE OF THESE MODELS, OR TEAMING UP WITH A COMPANY WHOSE EXPERTISE LIES IN BUILDING THESE VERY SAME MODELS AND/OR PANELS?

We have studied this option very carefully and we have highly skilled people in our team who were estimating missing data in the domestic markets prior to joining us.

 

We realised that even if the process sounds very similar, in reality it is a totally different business. Firstly, in the domestic markets retailers have less risk sharing their data – in most cases they are not in the contract sensitive concession business. With more sell-out data from retailers, the models become less complex.

 

Secondly, if, in a given domestic market, there is too much missing sell-out data, you have tools to estimate this, such as building a household panel [this is a panel in which selected families record all their purchases-Ed].

 

These panels need to be representative of the market in which you are estimating the missing data. Imagine the complexity of a worldwide representative panel of travellers?

 

So overall, it would therefore be naive to assume a domestic model would work in travel retail.

 

 

 

 

SO DO SUCH DIFFICULTIES AS THESE MAKE THE TRAVEL RETAIL INDUSTRY EVEN MORE DEPENDENT ON THE WILLINGNESS OF THE RETAILERS THEMSELVES TO SHARE DATA?

In a sense, yes. As I said earlier, if we could reach plus-60% of sell-out data from retailers, we could continue using our current highly complex formulas especially developed for the travel retail channel, but it would be more precise due to less missing data.

 

We would also be less dependent on suppliers (brands) providing us with their data. This would be a big relief to everyone. So believe me, we are trying very hard and we are constantly in touch with retailers and airports to raise that number who participate.

 

 

 

 

THIS IS NOTHING NEW THOUGH, AS I RECALL THE COMPANY HAS BEEN TRYING TO ACHIEVE THIS FOR THE PAST 30 YEARS. PRESUMABLY IT HAS GOT EASIER OVER THIS TIME?

Absolutely, after 30 years of building relationships of trust. This also shows two things. Firstly, it is not easy, because believe me we have been working on this relentlessly. To have the 40% we have now represents countless years and months of negotiations and problem solving [Lindh added that sometimes the willingness to share is there, but the data is not consolidated).

 

Secondly, we see the critical importance of having such a complex estimation model, because our industry is desperate for knowledge of where we are heading – even ironically, among those who don´t support us with data.

 

Imagine if we stopped being transparent and no longer communicated the data through TFWA and the DFWC with some of our top-line data – who would know where our industry is heading?

 

 

 

 

OK, SO YOU ARE TEAMING UP WITH M1ND-SET AND YOU JUST ANNOUNCED THAT YOU ARE HIRING ‘A WORLD-CLASS ACADEMIC’ TO WORK ON YOUR DATA MODELLING AND ANALYSIS. WHY ARE YOU DOING THIS AND WHAT ARE THE BENEFITS IN YOUR VIEW?

We really want to work hard on the first two processes I mentioned earlier. Firstly, we want more retailers to share data with us and I think the partnership with m1nd-set gives us much more to share with retailers in exchange for their sell-out data.

 

It also saves time for all our other partners; they get more time to analyse data instead of collecting and providing the data to us. Secondly, with m1nd-set onboard we have additional sources of data.

 

They interview over 100,000 passengers per year with travel retail-related questions and are the experts in the industry overall and in each of the categories due to all their various clients.

They also have teams of interviewers in plus-50 airports that can help with our side of the business too.

 

 

 

 

Thanks to m1nd-set’s exclusive access to direct airline bookings through their partnership with the relevant airline associations, we have access to the most precise air traffic and forecasting data.

 

For us it’s a great opportunity to further develop and enhance our model. As you know, we asked Prof. Seret to help us, because he is one of the world’s leading experts in the field of estimating missing data in ultra-complex environments.

He has consulted a range of industry leaders around the world and will help us to continuously improve our products and services.

 

SO ARE WE GOING TO BE LOOKING AT A MUCH BETTER RESOURCED GENERATION IN FUTURE WITH EVEN GREATER ACCURACY IN ITS BRAND AND RETAIL SALES RESULTS AND FORECASTING?

Well it is pretty good today, but the answer to that would have to be yes.