pacman::p_load(tidyverse, lubridate, janitor, DT, pander, plotly, scales)
# Loading in dataset
fs_data <- read_csv("C:/Users/felix/OneDrive/Documents/Family Search BI Analyst Interview/April_2025_Case_Study_Data.csv") %>%
clean_names() %>%
mutate(across(contains("_date"), ymd))This case study explores how new users interact with FamilySearch’s core features and redefines user segments based on depth and type of engagement. We segmented users into four behavior-based profiles based on what they actually did inside the platform - from browsing to building.
By rethinking our segmentation model, we gain sharper insight into where users thrive, where they stall, and how we might turn intention into lasting contribution.
engagement_summary <- fs_data %>%
select(
total_days_logging_in,
total_days_viewing_records,
total_days_adding_names_to_tree,
names_added_to_tree,
total_days_editing_tree,
total_days_uploading_memories,
memories_uploaded,
total_days_attaching_sources_to_tree,
sources_attached_to_tree,
total_days_using_get_involved
) %>%
summary()
pander(engagement_summary, caption = "Engagement Activity Summary Across Key Features")| total_days_logging_in | total_days_viewing_records |
|---|---|
| Min. : 0.000 | Min. : 0.0000 |
| 1st Qu.: 1.000 | 1st Qu.: 0.0000 |
| Median : 1.000 | Median : 0.0000 |
| Mean : 2.594 | Mean : 0.7469 |
| 3rd Qu.: 2.000 | 3rd Qu.: 1.0000 |
| Max. :272.000 | Max. :254.0000 |
| total_days_adding_names_to_tree | names_added_to_tree |
|---|---|
| Min. : 0.0000 | Min. : 0.000 |
| 1st Qu.: 0.0000 | 1st Qu.: 0.000 |
| Median : 1.0000 | Median : 2.000 |
| Mean : 0.7737 | Mean : 4.571 |
| 3rd Qu.: 1.0000 | 3rd Qu.: 5.000 |
| Max. :255.0000 | Max. :25432.000 |
| total_days_editing_tree | total_days_uploading_memories | memories_uploaded |
|---|---|---|
| Min. : 0.0000 | Min. : 0.00000 | Min. : 0.000 |
| 1st Qu.: 0.0000 | 1st Qu.: 0.00000 | 1st Qu.: 0.000 |
| Median : 0.0000 | Median : 0.00000 | Median : 0.000 |
| Mean : 0.7738 | Mean : 0.05084 | Mean : 0.246 |
| 3rd Qu.: 1.0000 | 3rd Qu.: 0.00000 | 3rd Qu.: 0.000 |
| Max. :262.0000 | Max. :166.00000 | Max. :5813.000 |
| total_days_attaching_sources_to_tree | sources_attached_to_tree |
|---|---|
| Min. : 0.0000 | Min. : 0.00 |
| 1st Qu.: 0.0000 | 1st Qu.: 0.00 |
| Median : 0.0000 | Median : 0.00 |
| Mean : 0.1961 | Mean : 7.47 |
| 3rd Qu.: 0.0000 | 3rd Qu.: 0.00 |
| Max. :255.0000 | Max. :235901.00 |
| total_days_using_get_involved |
|---|
| Min. : 0.00000 |
| 1st Qu.: 0.00000 |
| Median : 0.00000 |
| Mean : 0.01648 |
| 3rd Qu.: 0.00000 |
| Max. :176.00000 |
The engagement snapshot shows us that while many users explore or browse the platform, only a smaller portion actively contribute - and even fewer become power users. It highlights a clear drop-off between passive interest and active participation, pointing to a big opportunity to turn more browsers into creators. That gap between browsers and creators suggests high interest but low activation - meaning users arrive curious but don’t always know how to contribute.
na_counts <- fs_data %>%
summarise(across(contains("_date"), ~ sum(is.na(.)))) %>%
pivot_longer(everything(), names_to = "Column", values_to = "Missing_Count")
datatable(
na_counts,
caption = "Missing Values in Key Engagement Dates",
options = list(pageLength = 5)
)fs_data <- fs_data %>%
mutate(
creator = names_added_to_tree > 0 | total_days_editing_tree > 0,
browser_only = total_days_logging_in > 0 & names_added_to_tree == 0 & total_days_editing_tree == 0,
dormant = total_days_logging_in == 0,
memory_uploader = memories_uploaded > 0,
researcher = sources_attached_to_tree > 0,
behavior_segment = case_when(
dormant ~ "Dormant",
creator & memory_uploader & researcher ~ "Power User",
creator ~ "Creator",
browser_only ~ "Browser Only",
memory_uploader | researcher ~ "Supportive Contributor",
TRUE ~ "Unclassified"
)
)We categorized users into four behavioral segments to reflect how they engage with FamilySearch. Power Users are highly active, using at least three core features like adding names, editing trees, uploading memories, or attaching sources. Creators contribute by adding names to the tree but use fewer advanced features. Browser Only users explore the site (logins, record views) without making contributions. Dormant users registered but never engaged. These segments help us understand where users drop off, who drives value, and where we can focus efforts to guide behavior and boost retention.
segment_distribution <- fs_data %>%
count(behavior_segment, sort = TRUE)
plot_ly(segment_distribution,
x = ~n,
y = ~behavior_segment,
type = "bar",
orientation = "h",
text = ~scales::comma(n),
textposition = "auto",
marker = list(color = "mediumseagreen")) %>%
layout(title = "User Distribution by Behavior Segment",
xaxis = list(title = ""),
yaxis = list(title = "
"))Insight: Most users fall into Dormant or Browser-Only segments. These users log in but don’t contribute - yet. There’s untapped potential in converting these passive users with guided nudges and quick wins. Meanwhile, Power Users are rare but influential - we should amplify their impact.
age_seg_summary <- fs_data %>%
count(age_group, behavior_segment) %>%
drop_na() %>%
ggplot(aes(x = age_group, y = n, fill = behavior_segment)) +
geom_col(position = "stack") +
scale_y_continuous(labels = comma) +
labs(
title = "Behavior Segment Breakdown by Age Group",
x = "Age Group", y = "User Count", fill = "Segment"
) +
theme_minimal(base_size = 13) +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
age_seg_summary
Insight: The 18–35 age group is driving most platform activity, including Creator and Power User roles. But Browser-Only users are present across all age groups - especially 50+. These users may benefit from simplified interfaces and value-focused onboarding.
country_counts <- fs_data %>%
count(country, sort = TRUE) %>%
slice_max(order_by = n, n = 10)
ggplot(country_counts, aes(x = reorder(country, n), y = n)) +
geom_col(fill = "seagreen") +
geom_text(aes(label = scales::comma(n)), hjust = -0.1) +
coord_flip() +
scale_y_continuous(labels = scales::comma) +
labs(title = "Top 10 Countries by User Count",
x = "Country", y = "User Count") +
theme_minimal()
Insight: Brazil and the U.S. are your largest user bases. Tailor re-engagement and onboarding flows in those countries first - they offer the most leverage.
top_country_names <- country_counts$country
country_seg <- fs_data %>%
filter(country %in% top_country_names) %>%
count(country, behavior_segment)
ggplot(country_seg, aes(x = behavior_segment, y = n, fill = country)) +
geom_col(position = "dodge") +
scale_y_continuous(labels = comma) +
labs(title = "Behavior Segments Across Top Countries",
x = "Behavior Segment", y = "User Count") +
theme_minimal(base_size = 12) +
theme(axis.text.x = element_text(angle = 20, hjust = 1))
Insight: Power Users cluster heavily in Brazil and the U.S. Localization matters here - even small regional adaptations may turn “Creators” into “Champions.”