Michigan Named a Top U.S. State for Valentine’s Proposals, New Data

New data has ranked Michigan as a top ten state for Valentine’s proposal likelihood this year, the ninth-highest state in the U.S. based on engagement ring shopping and proposal-planning search behavior ahead of February, according to a new nationwide Valentine’s Proposal Likelihood Index.

Analysis from wedding experts at Walters Wedding Estates found that Michigan over-indexes for engagement-related search activity linked to February proposals, ranking behind California and ahead of Washington, with an index score of 162, much higher than the national average of 100.

The index is based on engagement ring shopping and proposal-planning searches captured during the peak November–January buying window, when most proposers purchase a ring ahead of Valentine’s.  According to The Knot, most proposers buy a ring four months or less before proposing, with 60% purchasing within two months, making November–January ring searches strong indicators of February and therefore, Valentine’s proposals.

Ring style and metal preferences were analyzed separately and combined to indicate the most likely engagement ring configuration by state. In Michigan, three-stone styles and gold are the most searched-for options, offering a snapshot of the kind of ring that could be appearing on fingers in Michigan this Valentine’s.

Full details below:

About Walters Wedding Estates:

Walters Wedding Estates is a wedding venue and event company founded in 2006 by Keith and Sarah Walters. With two decades of experience helping couples tie the knot and a team of over 750 members, the company has grown from a single property into a portfolio of more than 40 distinctive wedding venues across Texas and Atlanta. Walters Wedding Estates offers all-inclusive packages and services designed to simplify planning and deliver beautiful, stress-free wedding experiences from first tour to final toast. 

You can read more about them here.

Research Objective: To identify which U.S. states are most likely to see higher levels of Valentine’s proposals using engagement-related search behaviour.

Methodology Summary 

Walters Wedding Estates developed a Valentine’s Proposal Likelihood Index by analyzing Google Search Trends, specifically Google Shopping data and web data across a total of 74 engagement-related keywords. The study combines high-intent engagement ring purchasing behaviour with proposal-planning searches to identify where proposal intent is most concentrated across the U.S.

The analysis uses a date range from 1 November 2025 to 15 January 2026, reflecting established engagement ring purchasing timelines. According to The Knot, most proposers buy a ring four months or less before proposing, with 60% purchasing within two months, making November–January ring searches strong indicators of February and Valentine’s Day proposals.

Each U.S. state was indexed against the national average based on engagement-related search activity. Purchase intent was weighted more heavily than planning behaviour to reflect real-world proposal likelihood. States were then ranked and categorised as high, average, or low likelihood.

Finally, ring style and colour preferences were analysed to predict what an engagement ring is most likely to look like in each state — offering a data-led snapshot of the kind of ring someone might be proposed with this Valentine’s.

Detailed Methodology

Data Sources & Date Range

To estimate which U.S. states are most likely to see the highest number of Valentine’s proposals, Walters Wedding analyzed Google Search Trends data, specifically Google Shopping and web search search interest across 74 engagement-related keywords. The analysis used a date range from 11 January 2025 to 15 January 2026, designed to capture the full planning and ring-purchasing window associated with Valentine’s proposals.

According to The Knot’s 2024 Jewelry and Engagement Study, most proposers purchase an engagement ring four months or less before proposing, with 60% buying within two months. As a result, ring purchases made from November onwards are strongly indicative of proposals occurring in February, including Valentine’s Day.

The analysis window extends into mid-January to capture late-stage purchasing and final planning behaviour immediately ahead of Valentine’s.

Keyword Framework

The keywords were grouped into three core analytical categories designed to capture different stages of proposal intent. Only keywords with sufficient and reliable search interest were included to ensure meaningful state-level comparisons.

1. Purchase Intent (Ring-Buying Behaviour)

This category captures users demonstrating intent to purchase an engagement ring, as they’re specifically searching via Google shopping. To ensure comprehensive coverage, we included high-volume variants that reflect common ways consumers search for rings.

Keywords included:

• Engagement ring
• Diamond ring
• Buy engagement ring
• Buy diamond ring
• Silver engagement ring
• Gold engagement ring
• Silver diamond ring
• Gold diamond ring 

2. Planning Confirmation (Proposal Intent)

This category captures users actively planning a proposal, indicating high likelihood of imminent engagement activity.

Keywords included:

• Proposal ideas
• How to propose
• How to propose to your girlfriend
• Proposal tips
• What to say when proposing
• Engagement proposal
• Wedding proposal ideas
• Romantic proposal ideas
• Best proposal ideas
• Proposal speech
• Planning to propose 

3. Ring Style Preference

To understand regional differences in ring style preferences, we analyzed searches for specific ring types. Each ring type was evaluated using the formats:

• “[Ring type] engagement ring”
• “[Ring type] diamond ring” 

This approach ensured consistent capture of search intent across styles.

Ring types analyzed (30 total):
Oval, pear, heart, trillion, baguette, radiant, cushion, asscher, emerald, princess, round, marquise, vintage, halo, three-stone, antique, solitaire, art deco, milgrain, modern, gemstone, pavé, bezel, classic, natural, prong, split shank, east-west, double band, lab-grown.

Some ring types (e.g. square engagement ring, channel engagement ring) were excluded due to insufficient regional search interest. All ring styles with meaningful state-level search interest were included.

Data Aggregation and Indexing

For each U.S. state, search interest was exported and aggregated separately for:

• Purchase intent keywords
• Planning confirmation keywords 

State-level totals were then converted into indices using the following formula:

Index score = (State total ÷ National average) × 100

An index score of 100 represents the national average. Scores above 100 indicate proportionally higher engagement-related search activity relative to the national baseline.

Valentine’s Proposal Likelihood Score

To produce a single, comparable score per state, the two indices were weighted based on their relative importance to real-world proposal likelihood.

• Purchase intent: 80% weighting
• Planning confirmation: 20% weighting 

This weighting reflects the belief that ring-purchasing behaviour is the strongest indicator of an imminent proposal. While the exact weighting reflects a reasoned modelling choice, the index is designed to highlight relative differences in engagement likelihood by state rather than produce a precise prediction.

Final Score Calculation:

• (0.8 × Purchase Intent Index)
• (0.2 × Planning Confirmation Index) 

Scores were rounded to the nearest whole number.

State Classification

States were ranked from highest to lowest based on their final index score and categorized as follows:

• High likelihood: 140+
• Average likelihood: 80–139
• Low likelihood: Below 80 

These thresholds reflect engagement-related search activity relative to the national average rather than an absolute maximum.

Ring Style and Color Analysis

Once proposal likelihood was established, regional ring preferences were incorporated for additional insight.

Ring Style

For each state, the ring type with the highest search interest over the time period was identified. In cases where multiple ring styles showed equal search interest, all were included in the final results.

Ring Colour

To estimate likely ring colour, we analyzed search interest for the following purchase-intent keywords:

• Silver engagement ring
• Silver diamond ring
• Gold engagement ring
• Gold diamond ring 

Search interest for silver-related terms and gold-related terms was summed at the state level. The higher total was used to indicate the most likely ring colour preference in that state. 

This approach allowed us to work around the limited search volume of ultra-specific queries (such as “gold cushion diamond ring”) by combining two key signals: ring style and ring colour. Together, these insights enabled us to predict what an engagement ring is most likely to look like in each state — offering a data-led glimpse into the kind of ring you might be proposed with this Valentine’s.

Valentine’s Engagement Likelihood Index

Restrictions & Acknowledgements

• Lower index scores do not indicate fewer engagements. They reflect lower volumes of engagement-related search activity, which can be influenced by population size, rurality, internet usage, and how people plan major purchases. The index is designed to highlight where proposal intent is most concentrated, not where proposals are not happening. 
• Purchase intent was weighted more heavily than planning behaviour in the final index, reflecting the stronger link between ring-search activity and imminent proposals. 
• Search interest is used as a proxy for consumer intent, reflecting relative likelihood rather than confirmed proposal activity. This analysis does not guarantee that individuals searching for engagement rings went on to make a purchase. 
• Google Trends data reflects relative search interest rather than absolute search volume and is influenced by population size and online behaviour. 
• The analysis does not account for offline purchasing behaviour or non-Google search activity, which may vary by region. 
• Ring style and metal preferences are inferred from aggregated search behaviour and should be interpreted as indicative trends rather than definitive predictions.