Buying a list with one filter is what everyone does, which is precisely why it does not work well. Stacking is the practice of layering conditions until the list is small enough that almost nobody else has assembled it.
It is the highest-leverage technique in this area and most investors have never deliberately done it.
The Arithmetic That Makes It Work
Stacking is the technique that separates a bought list from an assembled one, and it underpins everything in lists and data for real estate investors.
Consider a county with twelve thousand absentee-owned properties. That is a list anyone can buy, and the response rate reflects how many people bought it.
Now layer conditions. Absentee, plus held more than ten years, plus an open code violation, plus a vacancy indicator. That might return two hundred records.
The instinct is that two hundred is too few to matter. The arithmetic says otherwise, because your cost is per piece rather than per record. Mailing two hundred records six times costs a fraction of mailing twelve thousand once, and the response rate on the stacked list is not comparable.
The other advantage is competitive. Anyone can pull absentee owners. Almost nobody assembles that specific combination, which means the owners on your stacked list are receiving your mail rather than eleven pieces.
What Each Layer Is Doing
Layers are not interchangeable. Each one is a proxy for a different kind of pressure.
Length of ownership. A proxy for equity and for fatigue. Someone who has held a rental fourteen years has both the ability to sell and, commonly, the accumulated exhaustion.
Absentee status. Distance makes management harder and attachment weaker.
Distance from the property. A stronger version. An owner two states away has a different relationship to the property than one across town.
Vacancy indicators. A property nobody is occupying is costing money and producing nothing.
Code violations. An event with a clock attached, and a cost the owner may not want to fund.
Tax delinquency. Financial pressure with a deadline.
Property condition or age. A proxy for deferred maintenance the owner may not be able to address.
Owner age, where available and used carefully. Life stage correlates with transitions, and this is territory where fair housing considerations apply and warrant care, per fair housing for investors.
Estate or probate indicators. A situation rather than a characteristic.
Situations Beat Characteristics
If you take one idea from this, take this one.
A characteristic is something ongoing and stable. Absentee ownership. Property type. Equity position. These describe a person's circumstances and say very little about whether this month is different from last month.
A situation is an event with a clock. A probate filing. A code violation notice. A tax delinquency. A divorce filing. An eviction.
Events indicate motivation because something has changed and there is pressure to respond. Characteristics indicate capacity to sell rather than reason to.
The strongest stacks combine both: a characteristic establishing that a sale is feasible, and an event establishing that now is different. Long-held absentee ownership plus a fresh code violation is exactly that shape.
How Far to Stack
There is a point where layers stop helping.
Too few layers produces a crowded list. Too many produces thirty records, which cannot support a campaign regardless of quality, and which may be small enough that the filters have selected for data errors rather than for reality.
The practical target for most markets is a list large enough to mail repeatedly for several months, which usually means a few hundred to a couple of thousand records depending on the county.
The way to find it is to add layers one at a time and watch the count. When it drops below what you can sustain a campaign against, remove the least predictive layer rather than the last one added.
Building a Stack Without a Vendor That Supports It
Most vendors offer some filters and not the combinations that matter, which is where the assembly work comes in.
The method is to pull each dataset separately and match them on the property address or parcel identifier.
An absentee list from a vendor or the assessor. A code violation list from municipal records. A tax delinquency list from the treasurer. Then match, and the intersection is your stack.
The matching is the friction, since addresses are formatted inconsistently across sources and parcel numbers are more reliable where available. Normalizing addresses before matching is most of the work, and it is exactly the kind of task that suits delegation or tooling, as in AI for data enrichment.
That friction is the point. A list requiring three sources and an afternoon of matching is a list your competitors did not build.
Testing Whether a Layer Earns Its Place
Layers are hypotheses and most investors never check them.
The method is to run two versions of a list that differ by one layer and compare, not on response rate but on conversations and eventually on deals. A layer that raises response and lowers conversion is selecting for curiosity rather than motivation.
At investor volumes this takes a season to read, which means testing one layer at a time and being patient, explored in why small sample marketing numbers mislead.
The layers that consistently earn their place across markets are the event-based ones. The characteristic-based ones vary more, which is why testing yours rather than importing someone else's stack is worth the time.
The Stack Nobody Builds
Available in most counties, and almost nobody assembles it.
Long-held, absentee, entity-owned, with an eviction filing in the last two years.
That combination describes a landlord who has held for a long time, does not live nearby, holds professionally enough to use an entity, and has recently been through the specific experience that makes landlords stop being landlords.
No vendor sells it, because it requires court records matched against ownership data. It is small, it is uncontested, and the people on it are unusually likely to take a call about selling, discussed in tired landlord leads.
Negative Layers
As valuable as the positive filters, and almost nobody applies them.
Stacking is usually described as adding conditions that indicate motivation. Removing conditions that indicate its absence works just as well and is easier, because the exclusions are unambiguous.
Exclude properties that sold in the last twelve months, since the owner just transacted. Exclude anything with a recent permit for substantial work, which signals investment rather than exit. Exclude owner-occupied properties if you work absentee niches. Exclude the price bands you do not buy in. Exclude anything already in your pipeline or on your suppression file, described in suppression lists.
Each exclusion costs nothing and removes records that could never have produced a deal. On a large pull the exclusions frequently remove a larger share than the positive filters kept.
Where Stacking Goes Wrong
Stacking on unreliable fields. If a data point is wrong a third of the time, layering on it removes good records and keeps bad ones.
Assuming intersection means motivation. Four conditions overlapping is a strong signal, not a certainty, and the conversation still does the qualifying.
Building a stack you cannot refresh. If assembling it took three days, doing it quarterly is a real commitment. Automate the pull where possible or accept a slower refresh cycle.
Over-narrowing into a list too small to test. Thirty records cannot tell you whether the stack works.
Stacking for Buyers, Not Just Sellers
The application nobody uses, and the buyer side is where it is easiest.
Cash purchases are recorded. So are the entities that made them. Which means you can assemble a list of everyone who bought a property without a mortgage in your county in the last two years, then layer.
Bought more than once. Bought in the price band you work. Bought property types you produce. Holds rather than resells, or resells quickly, which distinguishes a landlord from a flipper and determines what you should send them.
That stack produces a buyer list of people who demonstrably transact, which is a materially better starting point than a registration form, per building a cash buyer list.
It is also uncontested in a way seller lists are not, because most wholesalers build buyer lists by asking people to sign up rather than by looking at who actually bought.
The First Stack to Build
If you build one this month, build the simplest useful version: your primary niche characteristic, plus one event-based layer, plus length of ownership.
Three layers is enough to leave the crowded pool without requiring an unmanageable assembly project, and it will produce a list materially different from anything your competitors bought.
Then mail it repeatedly rather than once, because the entire economic case for a small precise list is that you can afford to contact it many times, which is the thing a twelve thousand record list makes impossible.