A Handful of Tips for Effective Demand Planning




Jean-Luc Hospital, Manager for Demand Planning Implementation at Merck Serono, joins Pharma IQ to discuss benefits and greater business impact of applying effective demand planning in the clinical supply chain.

Pharma IQ: Great of you to join us. Now, you're be delivering a presentation on the conference on the Benefits and Greater Business Impacts of Applying Effective Demand Planning in the Clinical Supply Chain. So Jean-Luc, how can we find a compromise between top-down and bottom-up approaches internally to ensure SOPs are aligned?

J L Hospital: Well, the key thing here is really to have the right person around the SOP table and to have the key role in responsibilities established before, of course. And the focus during the SOP should be put at the right product level and that grid of course between the participants and with the right unit of measure and understandable with everybody. In fact, I know we have a group of persons who are driving those SOPs for each global BU and preparing the meeting in advance. And, in fact, those meetings are held on a quarterly basis so the top-down view is presented which corresponds to the global business unit view of how the market will be developed and another view corresponding to the bottom-up view of our affiliates is confronted to the view.

So very visually, in fact, we can see the big gaps between the top-down and bottom-up approach and with good preparation one can highlight basically those major points and prepare decisions to be taken during those SOP to get aligned. So either the Global Business Unit would revise the whole number or sometimes it’s the market would have to change their number, or sometimes it’s also both of them who need to do their homework. Actually minutes of the meetings are shared among the participants and distributed at other meetings for actions and I can tell you that in Merck Serono in fact we had quite a few illustrative examples where without the two views wrong decisions would have been taken.

Pharma IQ: Thanks. And what are the key considerations to reinforce demand forecasting by reconciling top-down and bottom-up views?

J L Hospital: Well, the reconciliation in effect is not something straightforward. First, you need to have the right data referential which are those in the forecasting tool to store all the SOP result access at the specific aggregated level which can then be used by the subsequent steps of the planning. But one bottom-up consideration is that the level discussed during the SOP might be slightly higher than the one you want to have or your subsequent planning steps. And this is where you need to reconcile down the numbers you get at this higher level. So, in that case, several strategies can be adopted so we can reconcile based on strategic forecasts or proportionally to the history or based on the existing forecast.



But care must be taken that across a product family the same strategy can apply and this is where often the program can start because then you might aggregate things which have nothing to do together, and so the best is to receive and discuss at the right aggregated level for the level of planning you want, but it’s not always something easy because the Global BU wouldn’t have the granularity that you want to have. So reconciliation in that case is key and you must spend enough time in validation for this to make it happen and to make it correct.

Pharma IQ: And leading on from that, Jean-Luc, how do you forecast at various aggregation levels and with common units of measures?

J L Hospital: In fact, we use one aggregation level where the results of our SOP views is stored and another one where we store our market forecast coming from the affiliates. But then, depending on the Global Business Unit, the level of a planning we address, we might use a different level of aggregation, so a product family, a brand, sometimes on the strength of [unclear], to be able to visualise the numbers properly and to be able to discuss with the person on something which is understandable by everybody. This means indeed that for each brand we have managed to find the common unit of measure, so this can be simply a total number of tablets, for instance, for some products or the number of syringes of cartridges, and this unit of measure can be often straightforwardly related to numbers of patients which is then the discussion with the BUs. So this is how we do it at Serono.

Pharma IQ: And what reference points can be provided to challenge and improve forecasts across functions?

J L Hospital: So usually we are using a statistical prediction based on an aggregated history, so this can give already a good idea of the trend and [unclear] volume, so this is often taken as our baseline. We sometimes also work with scenarios, so optimistic or pessimistic, given by the Global BU. And in most of the cases we try of course to see at the same time the Global BUs’ aggregated local view and the statistical view. So let’s say we don't have only one point of reference but several and we try to compare them. The difficulty of course occurs when we start forecasting new products from scratch because unfortunately we don't have a magic ball in that case, and it makes things more difficult. But then our baseline would be more the work of the Business Intelligence Group who develop a simple forecast model based on the [unclear] of the drug and expected patients treated.

Pharma IQ: And finally what planning tools can be used to identify gaps and provide automatic alerts?

J L Hospital: Well, in our forecasting tool that we are using at Merck Serono, so it’s a GDA, we have alerts embedded in the user interface where the market affiliates are entering the forecast, and so they highlight, for instance any big variation between their estimation of the market demand and what we reckon as being a statistical forecast for them. But we've developed also internally some tools which are Excel, Access with links to Oracle, basically simple table, which allows us to make various reports so with actually reports for local usage where the granularity is, for instance, the country and the brand or country and SKU, so we check, for instance, at local level, the compilation or the forecast to cover our forecast horizon, which is key when you want to compare the global view and the aggregated local view.

We check also the market forecast with the statistical forecast over the next 12 months and for products which normally display a continuous historical pattern we make such check, so we can challenge basically the market on this. We check also the viability of demand from one cycle to another over a specific timeframe, and at global level also we are more analysing how the forecast is evolving, the forecast cycle of brands for specific time periods and across all countries. So this can give us a really good hint for when something is happening in the market, knowing that we have a lot of SKUs to handle, more than 14,000 SKUs, which is a lot for a small team as we are today in Merck Serono. So those kinds of alerts are working, but exceptions are really key for us.

So, in fact, we don't have really automated alerts, for instance, by email, so we check all the alerts that are operated by those tools before sending any queries or enquiries to our markets.

Pharma IQ: Thanks very much, Jean-Luc. And what are you looking forward to most at the event?

J L Hospital: Well, in this kind of event I like to meet of course a lot of persons with the same concerns as I and to see how they are able to develop their solutions and their way in our [unclear] world, so always very enriching and I expect that I can learn a lot from the others and I can meet interesting persons as well.

Pharma IQ: Sure. Well, Jean-Luc, thanks very much for your time today. It’s been great to speak to you ahead of the conference and we look forward to hearing more from you in Amsterdam in May.

J L Hospital: Okay, it was a pleasure, thank you.

IQPC

Please note that we do all we can to ensure accuracy within the translation to word of audio interviews but that errors may still understandably occur in some cases. If you believe that a serious inaccuracy has been made within the text, please contact +44 (0) 207 368 9425 or email helen.winsor@iqpc.co.uk.