How do you handle demand forecasting for new products or services?

During the development of a new product, it is typically about finding one with similar previous characteristics in terms of attributes: introduction price, launch period, level of investment in advertising, consumer group, territory, channel, durability, etc. Based on the above, we can select some previous demand curves with a lot of focus on aspects of inventory availability in the channel prior to the official launch, this will allow a comprehensive reading of the sales behavior and the required replenishment levels. Finally, as the S&OP team, a commercial decision will be made that allows a correct balance between demand and the desired inventory at the points of sale. I always advise measuring the trend in the first weeks.

Evalúa a tus proveedores y fabricantes: Al prever la demanda de un nuevo SKU, necesitarás evaluar la capacidad de tus proveedores y fabricantes. Establece un plan de distribución: Considera cómo planeas distribuir el producto y cómo esto podría afectar la demanda. Prepárate para la canibalización: Si tu nuevo producto es similar a uno existente, podría canibalizar las ventas del producto existente. Previsión del volumen inicial de ventas: Utiliza la investigación de mercado, las percepciones de los clientes y los productos comparables para hacer una estimación. Observa las ventas y los comentarios de los clientes: Una vez que el producto esté en el mercado, utiliza las ventas y los comentarios de los clientes para ajustar previsiones.

In a new product development stage , you just create the product and send as a sample to float into the market to test ,whether this has been accepted or rejected by consumers.

Handling demand forecasting for new products involves several steps. First, analyze historical data if available, considering market trends, seasonality, and external factors. Conduct market research, surveys, and focus groups to gather insights into customer preferences and potential demand. Utilize statistical methods like time-series analysis or regression models to forecast future demand based on historic patterns and market dynamics. Additionally, collaborate with sales, marketing, and product development teams to incorporate their input and expertise into the forecasting process. Continuously monitor and adjust forecasts as new data becomes available, remaining flexible to accommodate changes in market conditions or consumer behavior.

La tarea principal de los modelos de Análisis de Regresión (AR) es anticiparse a los hechos para prevenir contingencias y no sólo tratar de adivinarlas. Aún no existen modelos precisos y confiables para se?alar cuándo y en qué lugar ocurrirá un hecho específico. Pero los AR ayudan a entender mejor la incertidumbre. Francis Galton (1822 – 1911) creó este concepto para el análisis y estudio de variables. Existen modelos de una o más variables, con el propósito principal es analizar el impacto que tienen ellas sobre la de interés principal. Desde mi punto de vista, es una de las mejores herramientas estadísticas para el estudio de la demanda de los productos nuevos.

On new product introductions, it's important to incorporate product lifecycle curves in addition to the seasonality curves. Product lifecycle curves are not cyclical, and show the pattern that time series follow following a launch in the early hype and maturity phases. Including this product lifecycle element will make the time series forecast more robust.

Monitorar os desvios entre a previs?o e o real, e compreender as causas, é essencial para que os próximos ajustes sejam coerentes e estejam mais próximos à realidade. As previs?es s?o o primeiro passo em um processo de gest?o da cadeia de aprovisionamento, portanto, influenciam no resultado global de uma organiza??o.

Existen muchas formas de generar el pronóstico de un producto nuevo, y el resultado de la precisión de dependerá de un sinfín de variables. Además, la incertidumbre crece, si quién tomará el riesgo de ese emprendimiento eres sólo tú. Le pondrás mucha emoción, dinero y esfuerzo para que el consumidor adquiera tu oferta, y muchas veces no lo lograrás. Y pensarás ?en qué me equivoqué o qué hice mal? Hace algunos a?os (1962) Everett Rogers explicó en su libro, "La Difusión de las Innovaciones", porqué algunos productos (o servicios) triunfan y otros no. Y los describe en los siguientes 5 elementos: 1. La INNOVACIóN. 2. Los AGENTES DE CAMBIO. 3. El CANAL DE COMUNICACIóN. 4. El TIEMPO, y 5. El SISTEMA SOCIAL. Nos falta mucho por aprender

Demand forecasting for an item/product or services comes from the trend line of already identifiable products in the market and this has not been for new products/services that is entering into the market.

An effective approach is to leverage the sales volumes of existing products to predict demand for the new ones, since new products often share common attributes with existing ones, such as colour, size, or flavour. Of course, this is useful only when there are no major alterations in marketing or distribution strategies. This forecasting model extracts product attributes from existing products and cross-learns them with all products in the same category, as well as other categories, possessing matching attributes to predict the demand for the new product.

From my experience the best working strategy to accurately forecast new product is to find one or more miror product to the new product and use à machine learning algorithm to forecast the aggregation of those miror product but relate to the attribute of the new product and is trending potential. For finding those miror product there's several way but reliable and transparent way is by using product attributes, NLP and price distance.

Unless it's an entirely new and extremely innovative product with no potential competitors, there are always similar products that can serve as a benchmark - use it as a starting point!