Context:

Signet Jewelers* aimed better to understand our users’ purchasing behaviors and expectations when it comes to shopping on social media platforms such as TikTok, Facebook, and Instagram, to use the insights to inform a strategy to go planning with Meta on implementing Signet branded shops and what products to prioritize.

*Signet Jewelers operates as the parent company of Kay Jewelers, Zales, Jared, Banter by Piercing Pagoda, Diamonds Direct, Blue Nile, JamesAllen.com, Rocksbox, Peoples Jewellers, H.Samuel and Ernest Jones.

Problem Statement:

To better understand signet shoppers’ behaviors and expectations when it comes to purchasing jewelry from social platforms; identifying an appropriate strategy for social media e-commerce selling

Goals:

  • Define customer segments and their shopping behavior on social media

  • Identify a strategy for Signet’s presence on social e-commerce

Team:

1 UX Researcher (myself)

Role & Collaboration:

Throughout this project, I assumed the role of sole UX researcher, designing the survey and conducting the statistical analysis for this research initiative

Timeline: 1 two-week sprint

Recruitment Criteria:

Participants were recruited based on specific criteria:

  • Users who have purchased jewelry within the past 6 months or intend to do so in the next 6 months.

  • Participants had previously shopped at Signet brands

  • Three geolocation demographics: United States, United Kingdon, and Canada

Code (Pyhton)

Methodology:

A survey was employed to quantify user sentiment toward social e-commerce:

Total sample size (N) = 166 :

  • US (n1) = 55

  • UK (n2) = 56

  • CA (n3) = 55

Survey Questions (Examples)

Predictor Question:

  • (Q2) On average, how much do you spend per session when shopping for jewelry online?

    • Less than $100 

    • $100 - $250 

    • $250 -500

    • $500 –1000 

    • $1000+ 

Target Questions:

  • (Q32) How much do you agree with this statement: “I would purchase engagement jewelry through social media.”?

  • (Q33) How much do you agree with this statement: “I would purchase everyday jewelry through social media.”?

  • (Q33) How much do you agree with this statement: “I would purchase special occasion jewelry through social media.”?

Statistical Analysis:

A Power analysis was conducted to determine the required sample size to observe a statistically significant effect yielding ~55 participants per group.

  • Effect size = 0.3 (Moderate effect size)

  • alpha = 0.05

  • # Significance level power = 0.8

Bootstrapping (resampling) was also implemented to better understand the range of outcomes because of a lack of resources to attain survey respondents (200+)

A Priori:

Null Hypothesis (H₀): There is no difference in the likelihood of purchasing engagement jewelry through social media between high-income shoppers and low-income shoppers

Alternative Hypothesis (H₁): High-income shoppers are more likely to purchase engagement jewelry through social media than low-income shoppers

Data Analysis & Post Hoc Tests

As a result of non-parametric (not standardly distributed) data, I utilized the Krukal-Wallis test to analyze the variance between user groups for the target data, then followed up with Dunn’s test to analyze which specific groups showed a statistically significant difference.

The data results were visualized in the form of a bar plot (see below),

Results

Significant result for Q2 and Q32: Test Statistic = 29.42, p-value = 6.43e-06

Significant pairwise comparison: $1000.00 (mean=4.36) vs Less than $100 (mean=2.07)

Dunn's d = 0.604, p-value = 4.50e-05

Bootstrapped Power with Increased Sample Size Kruskal-Wallis: 1.00 Dunn's Test: 0.96

Shoppers who spend 1K on average on jewelry purchases are likelier to purchase engagement jewelry through social media than those who spend less than $100 on average. This follows a common sense understanding of shopping for jewelry, particularly engagement jewelry. This could be due to the normative expectations associated with engagement jewelry such as having a high price tag. According to our other survey People expect less expensive items when shopping on socials.

Concerning the likelihood of purchasing jewelry, engagement jewelry yields the lowest mean compared to everyday jewelry and special occasion jewelry.

Full Takeaways and Recommendations

  • Price is an important determining factor for most shoppers when shopping for jewelry and determining payment methods. This sentiment matches common sense intuition

  • Our shoppers are willing to buy jewelry on socials. However, there is a discrepancy when it comes to social media on desktop vs mobile 

    • Shoppers on social media for desktops are less willing to buy jewelry than if they were on mobile

  • Shoppers are willing to buy jewelry, of any type, through social media, but mostly Everyday Jewelry

    • Shoppers are less likely to purchase 1) engagement jewelry on social media than 2) everyday jewelry or 3) special occasion jewelry

  • A user’s decision to move forward toward the traditional e-commerce website after starting shopping on social media is likely circumstantial

    •  For Instagram, shoppers are more likely to stay on Instagram  for a purchase than switch from Instagram to the brand website

    • Shoppers believe shopping on a brand's website will be more expensive or the same price as shopping through social media (up to 2x in price).

      • This may be due to expectations of either sales/discounts on Instagram or overall less expensive items presented on the Instagram shop

Immediate Recommendation:

Given the opportunity to host products on meta platform stores (particularly Instagram for mobile), Signet should prioritize items in the range people are willing to spend (up to $1000). This will likely consist of everyday and special occasion jewelry but exclude engagement/wedding jewelry.

Limitations and future research:

Due to restricted budget and the nature of non-parametric tests, not every segment had a representative sample (for example: dollar amount users are willing to spend above 2500)

In future iterations, it may be worth sending the survey out to more customers. In addition, there is some insight gathered that could benefit from more context, so it’d be worth following the research up with user interviews on topics such as 1) why people may be less likely to buy engagement jewelry online compared to other types and 2) why exactly people who spend less on orders are potentially less willing to shop for engagement

Next

Guided Checkout