cover article data 2.0 fashion data

Data Quality, the missing link in retail 2.0

Data qualitymanagement is key to the success of the data-driven transformation of retail companies.

Is the influx of data making your data strategy unmanageable? Data quality is there to put everything in order. Since the advent of Big Data, retail leaders have been looking for levers to optimise their performance but have struggled to recup the fruits of their investment, especially when it comes to Data & AI projects.

Not all data come from the same source: some are correct, others are unusable. This is often thought to be due to the inaccuracy of the data collected, whereas these disparities are linked to poor data management.

Today, every company is aware that data is an essential commodity for their development, but without a data-centric strategy, they cannot implement a viable activity for their business model. However, few of them are also aware that it is not enough to extract data in order to achieve this, it is above all necessary to have quality data.

According to research firm Gartner, a company with poor quality data can make up to €12 million ($15 million) in losses. This loss is compounded by other difficult challenges: activating additional levers to compensate for the lack of quality, making the data processing process more complex by adding correction steps, etc., leading to new expenses and putting a brake on the strategy.

Collecting poor quality data has a negative impact. Many companies have already realised this, and it has given rise to the term “Garbage In, Garbage Out”: if data is collected without effective data quality management, the decisions made on the basis of this data will be distorted or even wrong. And in the age of Big Data, where billions of data are extracted every second, a quality-oriented data management strategy is needed.

CRITERIA FOR DEFINING DATA QUALITY

Data Quality allows us to measure the state of the data collected and to improve it before it is analysed.Its objective is to guarantee the reliability and veracity of the data, two dimensions that are essential for the accuracy of business reporting.

There are several criteria for defining data quality: these indicators are common to all domains but their scope is particularly relevant to the new world of retail.

Completeness concerns not only the data itself, but also all the information relating to the activity on which it depends. For example, in the case of in-store sales, receiving data only on transactions will not provide a complete analysis of business activity. The solution would be to send, for example, information on the delivery schedule or the location of products sold in the shop, or even a repository of the shop layout or potential turnover per area.

Another important indicator in quality management is the freshness of the data. This means whether the data are recent and well updated. The level of freshness varies according to the temporality of the statistics but also according to the use cases. This is a criterion not to be neglected, given the current situation of containment, customer experience trends and consumption habits have been drastically disrupted, which necessitates having the first hand knowledge of the data in order to derive informed analysis.

The degree of precision is a central element in attesting whether the information received is qualitative or not. As with completeness, accuracy is not only about the value but also about its context. Thus, the characteristics of a product need to be detailed, with a fine granularity, and including additional information on its composition. It would make sense to combine these products with other complementary products. An underdeveloped data strategy may alter the accuracy of the data, which will be perceived as erroneous.

In order to implement an omnichannel strategy, retailers need to make purchase and inventory data available to all stakeholders in the customer experience.
The time has come to focus on phygital commerce (the fusion of physical and digital) and to take into account the fact that the buying process takes place on different sales channels. Also, the management of data availability must not be neglected: all sensitive information, especially customer data, cannot be accessible to everyone for compliance reasons.

SOURCES OF ERROR

One of the best known advantages of data mining is that the values are necessarily accurate. Although this is a common idea, it is often far from reality! The accuracy of the information gleaned depends on many factors, a simple human error such as a typing mistake or a technical incident such as a cash register failure can jeopardise the authenticity of the data. These erroneous values aren’t automatically detected, hence the need to rely on a monitoring system to detect inaccurate data and resolve them in time.

Data coherence and consistency are intrinsic to database management, and aggregated information can end up with an alternative value to the original one. For example, the SKU of a product (Stock Keeping Unit) displayed in a sales receipt may have a different length value than the SKU affiliated to the same product in its repository. In this case, the standardisation of values helps to anticipate this type of problem, which prevents the tracking of the product life cycle.

These concepts complement the dimensions of data quality management. But before embarking on a Data Quality Management process, it’s necessary to ensure that all teams, both business and technical, are aware of data quality management.

It is often thought that data quality is lacking for technical reasons. Sometimes we even blame human error, which paradoxically makes it difficult for decision-makers to trust the analysis made of their data and prefer to rely on their experience or even their commercial instinct, rather than on factual reports. Using data to make business decisions is not enough. Data is complementary to human decisions, but having it ensures that you are on the right track and moving towards growth.

The data collected loses quality for technical and human reasons, but another factor that is underestimated by the various businesses is that the quality of the data stems from its organisation. For an optimal organisation, it is necessary to communicate at all levels: transparently, with a standardisation and a specific vocabulary, defined by all the people in charge of the exploitation of the data. At the height of data governance, Data Quality is everyone’s effort.
Data governance provides new rules for certifying the reliability of data, which is essential for data quality management. By defining a data quality strategy with data governance management, you ensure two fundamental aspects: data security and compliance with the General Data Protection Regulation.

Implementing a Data Quality strategy is key to meeting today’s retail challenges. The data extracted concerns the entire value chain and even beyond, as any activity generates data that is ready to be exploited. Without this, the analysis and insights derived from this data will not be relevant. Data Quality is the key to harnessing data effectively and in compliance. It guarantees a CRM strategy adapted to the expectations of consumers and in compliance with the RGPD legislation.

Tailer Logo

a Data Platform, adapted to the needs of retailers, that supports them in their Data-Driven transformation

Did you like this article ?

Do not hesitate to share

Read also

We are on Twitter

You currently have access to a subset of Twitter API v2 endpoints and limited v1.1 endpoints (e.g. media post, oauth) only. If you need access to this endpoint, you may need a different access level. You can learn more here: https://developer.twitter.com/en/portal/product

Tailer the platform retail data

We you accompany towards a new Fashion Business Zero Waste