Prior to the cooperation between Lisuto and Bic Camera, almost none of the Bic Camera’s products on the Rakuten platform were tagged, resulting in a considerable product discoverability weakness on the platform. Lisuto was tasked to solve the discoverability problem by tagging over 800K products in a wide scope of categories to the maximum possible extent as well as to provide the maintenance tagging for the new inventory. 3 month after the deployment of the project Lisuto followed up with a case study to determine the impact of data tagging on various business KPIs with focus on 3 key metrics - sales, page views and number of unique users.
The results of the study confirm that tagged products perform significantly better in terms of sales, page views and unique user visitation, compared to untagged products during the same period of time.
This case study is consistent with the previous studies on Ryuryu, Shirohato , York and others and clearly shows how tagging contributes significantly to improved performance, compared to untagged products. Tagged products showed significant improvement in all categories over the test period, contrasted with the uneven performance of untagged products.
Methodology
The case study has been conducted over a period of four months, in order to obtain a sufficient sample size. The first month, prior to tagging, served as the benchmark, and three months after the products have been tagged. 1000 top products with most page views were assessed in the study. The test period began in June 2020 and ended in September 2020.
Lisuto used the top viewed untagged products as a control group throughout the test period.
The Study
Sales
The analysis of results of the total sales shows the twofold advantage of the tagged products’ over untagged products. While the untagged products saw a significant drop in sales, the sales of the tagged products has actually increased.
In August and September, the sales of both tagged and untagged products have decreased. But in both months, the tagged products were much more resilient, and decreased by significantly less than the untagged products. Moreover, In September, untagged products continued their trend of losing sales, while tagged products showed a recovery.
Page Views
The page views statistics show a similar trend. Once again, July saw tagged products outperform untagged products in terms of page views, increasing the amount of page views over the month, while the statistics for untagged products dropped significantly.
In the next two months the website experienced reduced traffic, but once more tagged products showed their resilience. They experienced only a slight decrease in August, and recovered well in September to lose almost no page views. Untagged products, on the other hand, continued the decline trend for these two months, decreasing in page views sharply.
Unique users
The results for the unique users numbers keep in line with the trend. In July, tagged products drew more unique users compared to June’s numbers, while the numbers for untagged products decreased quite significantly.
In August, both types of products sustained a decrease in unique users numbers. Tagged products withstood it well, while untagged products numbers were worse than the month before. By September, the numbers for the tagged products bounced right back to their original numbers. Untagged products, however, continued their plummet.
Conclusions
This case study clearly demonstrates the major impact of tagging on sales and other key metrics for online retailers. Tagged products consistently and significantly outperformed untagged products, throughout the test period. Tagged products not only improved in all categories in the second month, but even when they have sustained a decrease in numbers, it was kept significantly smaller than the decrease sustained by untagged products. Tagged products showed their resilience even in times of lower website traffic.