Online sellers face significant challenges due to inconsistent data and fluctuating demand, making it difficult to implement accurate forecasting. According to Alexey Spas, CEO and founder of Instinctools, a software engineering company, traditional forecasting methods struggle in today’s omnichannel retail environment because they are designed for a stable, linear retail model that no longer exists.
Recent research shows that marketplaces accounted for 61 percent of total ecommerce GMV in Europe in 2025. With 47 percent of consumers starting their product search on marketplaces, many online retailers are active on these platforms. New marketplaces, such as the recently launched Argos marketplace, continue to emerge.
A report in 2025 indicated that most online sellers are active on six marketplaces, making it complex to keep track of demand, stock levels, and sales. Alexey Spas notes that retailers today are dealing with a complex set of market conditions where traditional growth tactics no longer suffice.
The complexities faced by these industries extend beyond simple logistics, into technological and data-related challenges. Data is frequently scattered across different teams or isolated systems, making it almost impossible to run a unified online strategy.
Related: Southern Europeans Top Chinese Platform Users
Traditional forecasting methods, such as spreadsheets, can become a bottleneck for online sellers, particularly when dealing with multiple SKUs, channels, markets, or warehouses. Manual consolidation or conflicting spreadsheet versions can also become an issue, leading to recurring stockouts or excess inventory.
Forecasting can take days, rather than hours, for online sellers. This is when retailers typically start looking for tools that can process larger, diverse datasets. AI or machine learning models can combine historical and real-time signals, detecting relationships and adapting to changing patterns.
These models can generate predictions on SKU-, location-, or channel-level. Instinctools’s software includes AI and ML facilities, although the company notes that human-led data engineering is still necessary. AI can significantly accelerate the preparation phase, as seen in one of their clients who implemented AI tools for data quality checks, reducing the time required by 60 percent.
As online sales continue to grow, retailers must adapt to the changing market conditions. With the rise of new marketplaces and the increasing complexity of online sales, traditional forecasting methods are no longer sufficient. Retailers must look to new technologies, such as AI and machine learning, to stay ahead in the market.
Related: London’s Legacy of Love: Tracing the History of Engagement Rings
The future of online sales will likely be shaped by the ability of retailers to adapt to changing market conditions and leverage new technologies. Alexey Spas notes that traditional growth tactics no longer suffice, and retailers must be willing to invest in new tools and strategies to remain competitive.
In the middle of this shift, retailers who fail to adapt will be left behind. The use of AI and machine learning can help retailers to better forecast demand and manage their inventory, but it will also require significant investments in data engineering and infrastructure.
For now, online sellers must use a combination of traditional forecasting methods and new technologies to stay ahead. The ecommerce industry is constantly evolving, and retailers must be willing to evolve with it.
