The use of AI is rapidly increasing in the process of planning and developing new products by companies. AI has already become a key decision-making tool in analyzing market trends, predicting customer needs, and optimizing product specifications.
However, there are clear reasons why many companies are not achieving the expected results even after introducing AI. The problem is data quality management.
The rapid changes in the market today make it more difficult than ever to plan new products. Consumers' minds change easily, technology is advancing rapidly, and competitors are constantly launching new products.
In this situation, companies are trying to reduce the development period for new products as much as possible.
It has become difficult to keep up with such rapid changes by conducting market research and making decisions in the same way as before. As a result, the number of new product failures is also increasing.
According to a recent survey by McKinsey, the failure rate of new products among consumer goods companies is 40% on average. This is 15% higher than it was 10 years ago.
In this situation, AI can be a very useful tool. This is because it can be used to quickly detect changes in the market and respond to them.
AI has brought about a revolutionary change in all stages of new product planning. Above all, it has made it possible to grasp market movements in real time.
Numerous pieces of information, such as social media, search data, and online reviews, are automatically analyzed by AI, allowing us to quickly capture changes in consumer needs.
In particular, the accuracy of predicting the market performance of new products has greatly improved. This is because AI comprehensively analyzes various data, such as what similar products have performed in the past, how the market has responded, and the competitive landscape.
Based on these analysis results, product planners can make decisions with more confidence.
In addition, the process of determining the detailed specifications of products has become much more sophisticated. AI analyzes in detail what features consumers prefer and how much they are willing to pay for each feature to find the optimal product configuration.
However, the performance of AI is entirely dependent on the quality of the data. No matter how sophisticated the AI algorithm is, if the quality of the input data is low, meaningful results cannot be obtained.
For example, global electronics manufacturer Company A introduced an advanced AI system to predict demand for new products, but the prediction accuracy was lower than that of existing statistical models due to inaccurate historical sales data and an unstandardized product classification system.
Market trend data is the starting point for new product planning. This includes conversations consumers share on social media, trends in search terms on major portal sites, and product reviews on online shopping malls.
This data helps to discover the true needs of consumers, which are difficult to identify through traditional surveys or market research.
When analyzing trend data, it is important to note that simply being mentioned a lot does not necessarily lead to actual purchases. The buzz online and actual market demand may differ.
Therefore, trend data should be analyzed in combination with sales data or market research results.
Competitor product data is key to determining our products' market positioning. We need to look at a wide range of things, including the specifications and prices of competitors' products, as well as their release dates, promotion strategies, and consumer reactions.
This allows us to find gaps in the market and our own unique selling points.
In particular, analyzing the launch patterns of competitors' new products can help predict future market conditions. Understanding the patterns of when major competitors usually launch new products, what features they emphasize, and how the price range is formed can provide useful insights for establishing our new product launch strategy.
Our historical product data is an important yardstick for gauging the success of new products. This includes not only sales and profitability by product, but also customer feedback, reasons for returns, and after-sales service data.
In particular, looking at how products in similar categories have performed in the past can help us more accurately predict the market potential of new products.
Particularly noteworthy for practitioners are the failures of products. By analyzing in detail why a product failed in the market and what consumers' complaints were, they can avoid repeating the same mistakes.
It is important to remember that there are as many insights to be gained from failures as there are from successes.
No matter how good a product idea is, it cannot exceed realistic constraints.
Supply chain data is essential for verifying the feasibility of product planning. You need to consider the price fluctuations of raw materials, the supply and demand of key parts, the operating rate of production facilities, and logistics costs.
In particular, the importance of this area is being emphasized more recently as the uncertainty of the global supply chain increases. You should also consider alternative options in advance in case the supply of certain parts or raw materials becomes difficult.
Based on this data, the cost of the product can be calculated accurately and the appropriate selling price can be determined.
IMPACTIVE AI systematically collects and analyzes various types of data to increase the success rate of new products.
The most basic data is the sales volume pattern data of past models. By analyzing the sales patterns of previous models, we can predict the sales volume of new models with a high degree of accuracy.
What is important here is not just the sales volume, but the 'pattern'. The key is to identify the sales patterns created by various factors such as seasonality, promotional effects, and the impact of competing products.
Consumer preference data is also very important. When deciding which features to include in a new product, you need to accurately identify consumer preferences. IMPACTIVE AI collects this information by crawling social media data or conducting surveys.
Even the best product is unlikely to succeed if the timing of its launch is not right, so the market environment at the time of launch is also very important. IMPACTIVE AI determines the optimal launch time by comprehensively analyzing market conditions at the time of launch, competitor trends, and the overall economic situation.
In addition, we also use external data, which amounts to around 50,000 pieces. We input the ERP data of our clients, various environmental data, and augmented/synthetic data into our deep learning model to improve the accuracy of our predictions.
Based on this vast amount of data, we predict the sales volume of new products in the first quarter, weekly sales volume, and total sales volume, and we also analyze the advertising intensity, sales volume by function, and changes according to the launch time.
This data-driven decision-making is actually showing great results. IMPACTIVE AI's predictive model shows an improvement in accuracy of 70-80% compared to the existing Excel-based demand forecasting and inventory management methods.
The most important thing in data quality management is standardization. If different departments interpret the same data differently or manage it in different formats, it can lead to major confusion later on.
For example, if the sales team manages sales volume in 'counts' and the production team manages it in 'boxes,' AI will inevitably produce incorrect results when analyzing this data.
In fact, one auto parts manufacturer was unable to perform proper analysis for a while even after introducing AI because the product codes used by each department were different. In the end, the company was able to properly utilize AI only after unifying product codes across the company and converting all past data into new codes.
Data standardization should start with basic information such as product name, model name, and specifications. In particular, data related to overseas customers or partners requires extra care as the language and unit system may differ.
The accuracy of data cannot be overemphasized. In particular, in new product planning, even a small error can have a big impact. This is because even a 10% misestimation of the market size can make or break a new product.
To improve data accuracy, a verification procedure should be established from the data input stage. It is recommended to set the range of input values in advance or build a system that automatically detects outliers.
For example, you can automatically check if the selling price is significantly outside the normal range or if the inventory amount is entered as negative.
Periodic data audits are also necessary. You should have a process in place to check the accuracy of key data on a quarterly basis and correct any issues immediately if they are found.
At this point, it is important not to simply fix what is wrong, but to identify the cause of the error and prevent it from recurring.
Data freshness is very important in AI analysis. Even the most accurate data may not be relevant to the current market situation if it is too old.
There should be a standard for updating market trend data in near real time, competitor information at least on a weekly basis, and company performance data on a monthly basis.
In particular, the timeliness of data has become more important as consumers' purchasing patterns have changed rapidly since the COVID-19 pandemic. Data from a year ago may not be enough to properly understand the current market situation.
Therefore, it is necessary to adjust the data collection and update cycle to the speed of market changes.
It is recommended to automate data updates. This is because manually updating data can lead to forgetting or delays. As much as possible, the system should be set up to automatically collect data, and the person responsible for manual input should be clearly designated to ensure accountability.
Data quality management for new product planning is not as difficult as you might think. The most important thing is to set clear standards before you start.
Many companies only realize the importance of data quality management after problems arise, but by then it is often too late.
First, let's take a look at the data collection stage. When collecting market data, you must check the reliability of the source. There is a lot of information floating around online, but not all of it is valuable.
In particular, data from social media and online communities can distort the actual market situation, so be careful.
When collecting information on competitors' products, continuous updates are key. If you continue to use the data you have collected once, there may be a big difference from the market situation. It is a good idea to create a process for regularly updating data on a weekly or monthly basis.
Accuracy is of the utmost importance when it comes to product data. In particular, sales and profitability data require close cooperation with the finance team, and special care must be taken to ensure that the data is not changed or omitted.
When collecting and managing data, you can improve the quality of your data by carefully checking the following items.
When collecting data, first clearly record the source of the data. This is useful for verifying the reliability of the data later or when an update is needed. It is also helpful to record the date of collection and the person in charge.
The data format must also be unified. For example, sales volume data must always be in units of 'number' and sales must be in units of 'one million won'. If the units are mixed, it can cause a great deal of confusion when analyzing the data later.
Handling missing values is also important. If there is no data, you need to decide in advance how to display it. Usually, it is displayed as 'N/A' or '0', but you need to establish a consistent standard within the team and follow it.
Finally, you need to set a data update cycle. You can set an appropriate cycle depending on the nature of the data, such as updating market trend data daily, competitor product information weekly, and your own performance data monthly.
Now, planning new products is no longer an era where you can rely solely on intuition and experience. It is essential to accurately listen to the voice of the market through data and predict the future with the analytical power of AI.
IMPACTIVE AI will be a strong partner for companies in this era of change, supporting more scientific and systematic new product planning.
Do you want to increase the success rate of new products? Do you want to stay one step ahead of market changes with data-based decision making?
With IMPACTIVE AI, you can. Experience IMPACTIVE AI's data-based new product demand forecasting service today.
.svg)
