BatchRank is BatchData's proprietary, AI-powered scoring model. It rates every U.S. residential property on how likely it is to sell in the next 6-12 months, so you can focus your time and budget on the owners most likely to move instead of marketing to everyone.
A standard property record tells you what a property is: its size, its owner, when it last sold. BatchRank adds a forward-looking signal: which properties are most likely to sell off-market or go on market (become actively listed) soon.
What you get with each score
Every scored property comes with three values. These terms are used consistently across BatchData:
Sale Propensity Score (0-100): how likely the property is to sell in the next 6-12 months, where higher means more likely and 100 is the highest. It is a relative ranking across all scored properties, so a score of 82 means the property ranks above 82% of scored properties, putting it in the top 18% of properties most likely to sell. It is not a literal "82% chance of selling."
Category (High / Medium / Low): a simple bucket for fast filtering and targeting, assigned from where the property's score falls in the overall ranking. Use the Category value, rather than a fixed number, when you describe or group results.
Status (Available / Not Applicable / To be Determined): whether a score exists. Available means the property is scored. Not Applicable means it does not meet scoring criteria, for example a non-residential property. To be Determined means the property is new to BatchData and has not been scored yet.
The essentials
Prediction window: the next 6-12 months.
Coverage: U.S. residential properties, nationwide.
Freshness: scores are refreshed weekly, and new or updated properties are rescored each week.
Access: BatchRank is a premium add-on. It is available through the BatchData API using the
batchrankdataset, and as a bulk dataset for platform partners.
What the score is built from
BatchRank draws on BatchData's property intelligence across eight broad areas and turns it into one score:
Investor and portfolio analysis: the owner's real-estate footprint nationwide, including number of properties owned, total value and equity, acquisition cadence, hold periods, and flip or landlord behavior.
Building permit and improvement activity: remodels, additions, roofing, solar, and kitchen and bath projects, with recency and job value, signaling reinvestment, life stage, or pre-sale prep.
Home value, equity, and market trends: the estimated value, equity percentage, and loan-to-value from the latest estimate, plus how value, assessment, and taxes move over time, with neighborhood context such as school district, HOA, and geography.
Mortgage and lien position: balances, rate, term, adjustable-rate (ARM) status, mortgage age, estimated payment, and refinance history across all lenders.
Ownership and transaction history: deeds, transfers, prior sales, and holding patterns.
Property distress signals: pre-foreclosure, liens, tax delinquency, and default notices.
Household demographics and life events: age, income, net worth, household composition, and recent life changes.
Property and occupancy detail: property characteristics, condition, owner-occupied versus absentee status, and vacancy.
The model is trained to take geographic market trends into account, so it adapts to local patterns rather than treating every market the same.
You do not need to be a data scientist to use it. BatchRank turns these signals into a single, actionable score for how likely a home is to sell.
How to search by BatchRank
BatchRank is built to stack with your other filters. Search a sale-propensity score range and combine it with filters like high equity, long ownership, or distress indicators to build a precise list.
Set the score range on searchCriteria.intel.salePropensity using its min and max values, then add any other filters you need:
{
"searchCriteria": {
"intel": {
"salePropensity": {
"min": 80,
"max": 100
}
}
}
}
The High / Medium / Low category is returned on each result, so you read it off the results rather than filtering on it directly.
Choosing a score threshold
BatchRank is a 0-100 ranking, so the practical question is where to set your cutoff. If you only contact properties scoring above a given threshold, you are trading precision against recall: precision is the share of contacted properties that actually transact, and recall is the share of eventual sellers your list captures. Which side matters more depends on your outreach economics.
When each contact is expensive (outbound calls, agent follow-up, in-person visits, concierge campaigns), precision matters most. Every wasted contact is staff time burned. Set a tight cutoff, for example a minimum score of 95 or 99, to maximize the share of contacted properties that actually transact, and accept that some eventual sellers will fall below your cutoff.
When each contact is cheap (postcards, direct-mail batches, broad email), recall matters most. The marginal cost of one more mailing is small, but every missed seller is a deal lost to a competitor. Set a looser cutoff, for example a minimum score of 70 or 80, to capture more of the eventual sellers, and accept a lower hit rate per contact.
Set the cutoff with the min value on searchCriteria.intel.salePropensity, and adjust it as you measure results.
Why it helps
Instead of buying a large list and marketing to everyone, BatchRank concentrates likely sellers into a far smaller, richer list, so your time and budget go where the sellers actually are. Tight score cutoffs give you the highest concentration of real sellers, and broader ranges capture more of the eventual sellers. You dial that tradeoff by choosing where you set the score range.
Common questions
How is BatchRank different from a standard lead list? A standard list gives you static facts about a property. BatchRank gives you a forward-looking signal about which properties are most likely to sell soon, whether off-market or by going on market, so you can act before a home hits the market.
What does the score mean? It is a relative ranking of how likely a property is to sell in the next 6-12 months. Higher scores mean higher likelihood, and a score of 82 ranks above 82% of all scored properties, putting it in the top 18% of properties most likely to sell. The High / Medium / Low category comes from the property's position in that ranking, so use the Category value rather than a fixed number.
What area does it cover? All U.S. residential properties, nationwide. The model is trained to take geographic market trends into account, so it adapts to local patterns rather than treating every market the same.
How often are scores updated? Weekly. New and updated properties are rescored each week.
How do I access it? BatchRank is a premium add-on, available through the BatchData API with the batchrank dataset and as a bulk dataset for platform partners. Contact our team to add it to your plan.