Properties
Last updated: November 19, 2026
Properties in Rally are the "columns" of your database that allow for powerful filtering, dynamic segmentation, and automated governance. They are “High-Signal Filters” rather than just storage. Every property you add should solve a specific problem you’ve faced in past recruitment or anticipate in future studies. If a data point doesn't help you include or exclude a participant with confidence, it may be noise.
Defining Properties for Rally
Guiding Principles
If you can't use it and rely on it to filter, segment, or personalize, don't store it.
Action-Oriented: Prevent the database from becoming a "data graveyard" by ensuring every property has a functional purpose in recruitment.
Focus on Utility: Reinforce that properties are the levers for finding the right participants and hydrating profiles with important qualifying data.
Promote Data Hygiene: Favor structured data (like dropdowns and dates) over messy, un-filterable "notes" or open text fields.
Leave Nothing to Interpretation: Where two properties could plausibly hold the same value, or a real-world value has no matching option, someone has to decide what to do — a teammate, or an AI acting on your behalf. Removing the ambiguity is more reliable than hoping the decision goes your way.
Key Properties by Common Recruitment Goals
Reduce frustration and frequency of recruiting by identifying the most common and useful data points about participants that should be captured in Rally.
Look ahead or look back at three research recruitment efforts. For each, ask:
Target participant: Who was the target?
Ex: Small business owners using our invoicing tool
Source of data: How did we find them?
Ex: Exported a list from the CRM of Small Business Owners and manually filtered by “Last Invoice Date”
Priority: What criteria was non-negotiable and what could be relaxed for a participant to be qualified?
Focus on properties that are key decision makers over nice to have information.
Frequency: How often is this data used to recruit for other studies?
Store properties that are more broadly applicable across studies.
Frustration: What was the most time-consuming part of finding these people?
Reduce frustration by including properties that are hard to collect but important to know
Guiding Questions for Properties in Rally
Use the following considerations to determine if the data should be stored in a property in Rally.
Inclusion Test: Will this property help me quickly identify a 'Must-Have' group for a study?
Exclusion Test: Will this property help me quickly remove people who would disqualify a study?
Personalization Test: Will this property help me customize my invitation to increase response rates?
One-Time vs. Long-Term Rule: Will this property be useful for a broader array of study recruitment needs?
Use a Property if the information is evergreen and will be useful for multiple studies (e.g., Industry, Plan Type).
Use a Screener Question if the information is specific to only one study (e.g., "What did you think of our social media post yesterday?").
Categorize Properties by "Utility"
Help uncover properties by breaking them into four functional buckets:
Identification & Communication
Without these, you cannot contact participants or schedule sessions.
Most of these are out-of-the-box properties in Rally and do not need custom properties to store and use.
Ex: First Name, Last Name, Email, Phone Number, Timezone, etc.
Demographics & Psychographics
Essential for ensuring a diverse and representative sample for your studies.
Many of these are self-reported data that’s collected via screener or sign-up form.
Ex: Job Title, Company Size, Industry, Age, Gender, Income, Location, etc.
Product & Behavioral Data
Allows researchers to target by behavioral analytics or other customer-specific metadata.
Most of these are stored in a CRM or analytics tool that must be synched with Rally or imported to stay current and relevant.
Ex: Account Plan (e.g., Enterprise vs. Free), Sign-up Date, Last Login, Features Used, NPS Score, etc.
Research Considerations
Crucial for governance and honoring participant needs. Preventing over-contacting and managing cooldown periods.
Many of these may be self-reported data that’s collected via screener or sign-up form.
Ex: Specialty Tags (e.g., "Accessibility User"), Research Preferences.
Choose the Right Property Type
Store the data in a format that helps with retrieval and minimizes confusion. Opt for structured formats where possible, choosing unstructured if the options are unpredictable.
In Rally, properties come in be stored in a few formats:
Property Type | Best for... | Example |
Single-select | Mutually exclusive categories (up to 100 items) | Account Tier, Job Role, Income Level, Age Range, Gender, etc. |
Multi-select | Overlapping traits or interests (up to 100 items) | Features Used, Products Used |
Date | Recency or lifecycle tracking | Last Login Date, Sign-up Date |
Number | Quantitative thresholds | Number of Seats, Total Spend |
Text | Strings often with unstructured or unpredictable values | Name, Email address, Phone Number |
Checkbox (T/F) | Simple binary flags | Has Opted In, Is Beta Tester, Has Registered |
Cover the States You Actually Track
For single-select and multi-select properties, the option set is a promise about what values are possible. A real value that has no matching option doesn't stay outside the property — it gets mapped to the nearest one, or dropped. Neither outcome announces itself.
Before you finalize an option set, list the values your source systems and your team actually produce, and check each one against the list. Pay particular attention to:
Stages added after the property was built: Properties that track a participant's status in a program are usually created with the two or three states that existed at the time. As the program matures and new stages appear, values arrive that the original option set never anticipated.
Values that exist in the source system but not in Rally: If a synced field carries five possible values and your property offers three, two of them have nowhere to go.
Deprecated options: Leave them in place if historical records still use them; removing an option doesn't reclassify the people who held it.
When you genuinely can't enumerate the values, a text property is the honest choice — accepting that it filters less well — rather than a select list that quietly rounds off the values it doesn't recognize.
Naming and Describing Properties
Two people should be able to look at your property list and reach the same conclusion about which property to use. So should the Rally Agent, MCP, or any AI tool your team connects. All of them should choose properties by considering their names, descriptions, and selectable options when finding and sizing audiences.
Naming
Use the term your organization actually uses, not a source system's field name. Account Plan beats acct_plan_c. If you can’t name the property with natural language, at least include this in the property’s description.
Make every name unambiguous on its own. If two properties could plausibly answer the same question, one is redundant or poorly named. This can often happen if you have data being synched from an analytics tool but also collect similar data as self-reported through surveys and screeners. Distinguish these by indicating which is self-reported and which is synched data.
Hold one convention for capitalization and spacing across the whole schema.
Avoid abbreviations only your data team would recognize. Consider how the entire organization would refer to that data, not just one team.
Descriptions
Write a description for every custom property. A good one answers three things in a sentence or two:
what the property means
where the value comes from
when a researcher should filter on it
Example: Account Plan — the customer's current subscription tier, synced daily from Salesforce. Use to target by segment; values are Free, Pro, and Enterprise.
Descriptions keep your team from guessing. An agent choosing between similar properties works from the same signals your team does — name, description, option set. A property whose meaning can’t be established from those is one it can’t confidently filter on.
Publish a property list. Once names are settled, distribute the list to everyone who writes research briefs or prepares participant files. Consistent property names in briefs and in spreadsheet headers are what make AI-assisted study setup land correctly the first time. Builders using the MCP can also look up property definitions themselves, so a clean, well-described schema is directly self-serve.
Review an existing schema. If your workspace has been in use for a while, perform an audit looking specifically for near-duplicates and orphans. Consolidating two overlapping properties is work you do once; living with them is a decision your team re-makes every time data comes in.
Bringing Data In Without Reshaping It
When an import file’s columns and values line up with your schema, an import is a straight transfer. When they don’t, new properties are only created on-the-fly if your role has the permission to do so. Creating a property as part of an import requires the the permission to Create and modify properties.
In the Rally interface, you map columns to existing properties yourself before the import processes. Only if you have permission to Create and modify properties will you be able to add new properties through this process.
Through the Rally Agent or MCP, importing a CSV or a pasted list auto-maps columns to existing properties if there is a strong match. A column for a property your schema doesn’t have is dropped or refused while the remaining data that matches updates. Only if you have permission to Create and modify properties will they be created on-the-fly through the import process if there is no match.
A property created during an import arrives with a name, a type, and a partial option set. That is the shortest path to a schema full of properties nobody can interpret later, including an agent.
Practices that keep the gap closed:
Distribute a sanctioned import template: One CSV with identity columns (email, first name, last name) plus your property columns as headers, and the exact option values as the accepted entries. Files built from it need no interpretation.
Create new properties before the import runs: If a file needs a field your schema doesn’t have, have a process for users to request a new property to be created first so it can be set correctly and associated with the right populations.
One practical note for any import: if the people imported are intended for a specific study, it is always best to import them directly into that study rather than into a population or database first, then searching for them afterwards.
Determine the Source of Truth
Review where this data will come from to ensure it stays fresh.
Automatic Sync: Use the Rally API or data system integrations (e.g. Salesforce, Snowflake, BigQuery) for key product and behavioral data.
Screener Forms: Use Rally Screeners, Surveys, or Sign-up Forms to capture self-reported data (e.g., "Why did you sign up?") and map those responses directly to properties.
If you use the Rally MCP to map screener or survey responses to properties after a study runs, it maps to properties that already exist.
Create the target properties, with their full option sets, before the study goes out.
Manual CSV Upload: For static lists from external tools.
Be careful about allowing a property to be updated by screener responses as well as automatic data syncs. Aside from contact information, consider having separate properties for self-reported data vs automatically synced data.
Example:
Products Used
Products the participant self-reports they use from your company and you want to hold on to and intend to have them update periodically.
[EXT SYS] Products Used
Products you have evidence that the participant has used based on analytics data in your internal systems and sync to Rally for recruiting.
Compliance
Before adding properties, verify if they need special treatment:
Is it Private? Mark sensitive data like PII or PHI (Personally Identifiable Information) correctly to protect privacy. This can be modified at any time.
Is it Unique? Email addresses always need to be unique to one person’s record. Other properties like user IDs or user tokens might also need to be unique. Mark these properties as unique when creating, this cannot be changed later or undone.
Syncing Properties with Data Integration
Rally offers the ability to use custom APIs or native integrations with other data systems (e.g. Salesforce, Snowflake, BigQuery).
➡ Bring Data into Rally
The most common and useful scenario is to bring data into Rally that helps achieve your recruitment efforts.
Make sure properties are created in Rally first and they match the property type and options from the source system.
If the type of data and options are too unpredictable to create as structured data in Rally, set them up as “text” properties and guide your team on how best to filter and search using text vs defined options.
Automatically sync data every 24 hours or trigger the sync when someone uses a CSV to import.
⬅ Update Your CRM
Avoid sending properties that hold self-reported data in Rally to your other internal data systems.
Only sync properties to your internal data systems that will change an action for your Sales or Success teams. If a piece of data won't help them have a better conversation or make a better decision, keep it in Rally to maintain a 'clean' CRM. Consider properties like:
Research Opt-in or Opt-out status
Last Research Participation Date
Last Research Contact Date
Lifetime Incentives Paid
👉 Rally Help Center: Connect Salesforce, Manage Salesforce
👉 Rally Help Center: Connect Snowflake, Manage Snowflake
Property Groups
Property groups allow you to organize your custom properties across the entire Rally People Database into logical groups that can be shown or not shown for one or more populations. Organizing properties into groups won’t impact filtering or searching, it simply makes locating and focusing on the right properties faster and easier for teams.
A property can only be associated with only one property group. By default, all custom properties live in a default “Custom Properties” group.
Any properties within a group that are restricted to specific populations will only be visible to team members who have access to those populations.
We recommend you organize properties into meaningful themes taking into consideration the purpose of the property and which populations the property is relevant.
Examples of property groups:
Demographics
Properties that describe statistical, socioeconomic, and personal characteristics of a population, such as age, gender, income, education, location, etc.
These properties are often relevant across all populations.
Product Data
Properties that might describe feature usage, days active, product tenure, login frequency, etc.
If you have different populations for different products in your database, this is a good opportunity to create multiple “Product Data” property groups that are specific to those populations and allow only the relevant populations to see them.
Product ABC Data is shared only with Population 1
Product XYZ Data is shared only with Population 2
Account Data
Properties that describe the participant’s account relationship including plan/tier, subscription details, customer success or account manager, etc.
It’s possible that these properties are relevant for only some populations.
Research Preferences
Properties to capture the participant’s preferences around research topics of interest, frequency of contact, research type preferences, etc.
These properties are often relevant across all populations.