Every Door Direct Mail vs. Data-Driven Mail for Roofing Contractors

Direct mail still works for roofing. It works because a roof is a problem people can see, a problem on every house, and a problem that arrives on a clock most homeowners ignore until it leaks. The argument inside the roofing world is not whether to mail. It is how to choose who gets the mail. On one side sits Every Door Direct Mail (EDDM), the United States Postal Service program that lets you blanket a carrier route for a few cents postage with no list at all. On the other sits data-driven mail: you build or buy a list, you filter it by property and homeowner attributes, you sometimes layer storm exposure and roof age on top, and you pay more per household to mail far fewer of them.

Most contractors pick one out of habit. The shop that grew up on storm chasing mails EDDM after every hailstorm because it is fast and brainless. The retail remodeler who hates wasting money mails a skinny targeted list and wonders why volume is thin. Both are leaving money on the table because they treated a tool like a religion. The right answer is route-by-route, season-by-season, and it depends on numbers you can actually calculate before you spend a dollar. What follows is the working model: the real costs, the response math, how to build a list that beats a blanket, where roof-age and storm data change the picture, the compliance lines you cannot cross when storm and insurance language enters the copy, and a decision framework you can run this week.

The two approaches, defined honestly

Before the spreadsheet, get the definitions straight, because half the bad decisions in roofing mail come from fuzzy terms.

Every Door Direct Mail (EDDM) is a USPS retail product. You pick carrier routes on the USPS mapping tool, you bundle your pieces in counts of 50 or 100, and the carrier drops one on every active delivery address on that route. There is no name, no address printing, and no list. You pay a flat per-piece postage rate (the EDDM retail rate, lower than First-Class) plus printing. You can filter routes coarsely by the demographics USPS attaches to a route (approximate median age, household size, income range), but the unit you buy is the whole route. If a route has 480 deliveries, you mail 480 pieces or none.

Data-driven mail (also called targeted, list-based, or saturation-minus mail) starts from a list of specific addresses you chose for a reason. The reasons stack: owner-occupied single-family homes, a property value band, an estimated home age, length of residence, an absentee-owner flag if you want rentals, and increasingly a roof-age range and a storm-exposure score per parcel. You print each piece addressed to the household, you pay First-Class or a presorted Standard/Marketing Mail rate, and you mail only the addresses that survive your filters. You might mail 40% of a neighborhood instead of 100%.

The distinction that matters is not "cheap vs. expensive." It is blanket vs. selection. EDDM buys you reach with zero waste-avoidance. Data-driven mail buys you the ability to skip the houses that will never call, at the cost of a higher price per surviving household and the work of building the list. Everything downstream flows from that trade.

There is one more practical wrinkle worth knowing before you map routes. EDDM comes in two flavors. EDDM Retail lets you mail up to 5,000 pieces per ZIP per day without a permit, paid for at the post office counter, and is the version most small shops use. EDDM BMEU (the higher-volume version that runs through a Business Mail Entry Unit) requires a permit and a mailing-agent setup but allows larger daily volumes and lets you add some simple addressing. For most roofing campaigns the retail version is plenty, and its 5,000-per-ZIP-per-day ceiling rarely binds because you are usually mailing a handful of routes at a time. The point to file away: "EDDM" is not one product, and your printer or mail house can tell you which lane fits your volume and whether they will drop it for you or hand you bundled pieces to take to the counter yourself.

What each one is genuinely good at

EDDM is good when the thing you are selling applies to nearly everyone on the route and the route itself is the filter. A brand-new shop introducing itself to a tight, uniform subdivision of 1990s-built homes after a confirmed hailstorm is a textbook EDDM case: almost every roof on that route is the same age, took the same hail, and is worth a knock. Paying to suppress a few houses would cost more than just mailing them.

Data-driven mail is good when the route is not the filter, which is most of the time. Mixed-age neighborhoods, scattered ownership, a city where the storm clipped six streets and missed twelve, a retail replacement offer that only makes sense to owner-occupants with 20-plus-year-old roofs, or a referral-style "we just finished three roofs on your street" piece. In all of those, the houses that matter are a minority of the route, and EDDM forces you to pay for the majority that will never respond.

The cost model: what a piece actually costs

Response rate gets all the attention, but cost-per-piece is where EDDM earns its reputation and where data-driven mail has to justify itself. Build the model with ranges, not single numbers, because printers and postage tiers vary by region and quantity.

Here is a representative cost breakdown per delivered piece for a standard glossy postcard (roughly 6 x 9 to 6 x 11, the EDDM-flat size). Treat these as planning figures, confirm current postage on the USPS site, and get live quotes from two printers before you commit.

The headline is real: EDDM is the cheapest way to put paper in a mailbox, mostly because you skip the list cost and ride the lower retail postage tier. A data-driven First-Class piece can cost noticeably more all-in once you add the list and addressed printing. That gap is the entire EDDM sales pitch.

But cost-per-piece is the wrong denominator. You do not buy pieces. You buy conversations, inspections, and signed contracts. The number that decides campaigns is cost per acquired job, and that number depends on response rate, set rate, close rate, and average job value, all of which move when you change who gets the mail.

Cost per piece is a trap; cost per job is the truth

Work a concrete example. Say you can spend a fixed budget on one neighborhood, and the neighborhood has 1,000 deliverable addresses. Assume a 30-year-old neighborhood where, realistically, maybe 350 of those homes have a roof old enough or storm-worn enough to be a live replacement prospect this year.

EDDM plan: mail all 1,000 at a low all-in cost per piece. You reach all 350 good prospects, plus 650 houses that are not in the market. Your response rate, measured against the whole 1,000, looks thin because the denominator is padded with people who have a five-year-old roof.

Data-driven plan: mail only the ~350 likely-due homes at a higher all-in cost per piece. You reach the same good prospects, skip almost all of the dead weight, and your response rate against the mailed quantity looks much stronger because the denominator is clean.

The question is which plan produces more jobs per dollar. That is decided by two things: how much more the targeted piece costs per household, and how accurately your data actually identifies the 350. If your data is sharp, you pay the premium on only the houses that can convert and you win on cost per job even though you lose on cost per piece. If your data is sloppy, you pay the premium and still mail junk, and EDDM wins. Data-driven mail is a bet on the quality of your selection. The rest of this is about making that bet a good one.

Response-rate reality for roofing mail

Ignore any vendor who quotes you a single magic response rate. Direct mail response is a distribution, and for roofing it swings on offer, timing, and targeting far more than on the paper. Industry-wide direct mail response figures published by the trade association for marketing professionals tend to land well under 1% for cold prospect lists and meaningfully higher for house lists (people you have touched before). Roofing has its own multipliers on top.

Think of response as a stack of multipliers rather than one number:

• Baseline interest. A cold roofing postcard to a random homeowner with no trigger event might pull a fraction of a percent. That is your floor.
• Roof-age relevance. Mail the same piece only to homes whose roofs are plausibly at end of life and the effective response climbs, because a larger share of recipients have the problem you solve.
• Storm timing. Mail into an area days after a verified, significant hail or wind event and response can jump several-fold for a short window, because the problem just became visible and urgent.
• Local proof. "We just replaced roofs at [three nearby addresses or a recognizable street]" outperforms a generic brand piece because it borrows trust from the neighborhood.
• Repetition. One drop underperforms a sequence. The same household seeing three touches over a few weeks responds at a higher cumulative rate than three different households seeing one touch each.

The practical takeaway: the way you lift response is by improving the denominator and the timing, not by chasing a clever headline. EDDM gives you reach and speed but cannot improve the denominator. Data-driven mail's whole job is to improve the denominator. So the strategic question becomes: in this specific neighborhood, this specific week, is the denominator improvable enough to pay for itself?

A worked response-and-cost comparison

Put numbers on it. These are illustrative planning numbers to show the structure of the decision, not guarantees; plug in your own.

Scenario: 1,000-home, 30-year-old subdivision, no recent storm, retail replacement offer. You estimate 350 homes are realistically in the market this year.

EDDM run (mail 1,000):

• Assume effective response among the 350 good homes is decent, but you also mail 650 homes where response is near zero.
• Suppose 7 calls come in total. Of those, 4 turn into inspections, 2 into signed jobs.
• Your total spend is 1,000 pieces at the low EDDM all-in rate.
• Cost per job = total EDDM spend divided by 2.

Data-driven run (mail the 350 likely-due homes):

• Same offer, but every piece lands on a plausible prospect, and you add a roof-age line to the copy that makes the piece feel personal.
• Suppose you still get 6 calls (you lost almost nothing by skipping the 650 dead homes), 4 inspections, 2 jobs.
• Your total spend is 350 pieces at the higher targeted all-in rate.
• Cost per job = total targeted spend divided by 2.

If 350 targeted pieces cost less in total than 1,000 EDDM pieces, data-driven wins outright, same jobs for less money. Often it is close, and then the tiebreakers are second-order but real: targeted mail produces a cleaner CRM (you know exactly who you mailed and can follow up), avoids annoying 650 uninterested neighbors with offers they will never use, and lets you reinvest the saved budget into a second or third touch on the good list, which is where the compounding response lives. EDDM wins when your prospect density is high (most of the route is in-market, like a uniform post-storm subdivision) or when your data is too weak to trust the selection.

Run this with your own per-piece costs and your own honest set/close rates. The math, not the brochure, tells you which tool fits the route.

The break-even formula, written out

If you want one equation to keep on your phone, here it is. Targeting beats blanket whenever the cost you save by not mailing dead houses exceeds the extra you pay per good house for selection. Spelled out:

Let P = number of likely-due homes on the route, D = number of dead (not-in-market) homes, c_eddm = all-in EDDM cost per piece, and c_target = all-in targeted cost per piece. Blanket spend is (P + D) x c_eddm. Targeted spend is P x c_target. Targeting is cheaper for the same jobs whenever:

P x c_target < (P + D) x c_eddm

Rearranged, targeting wins when the premium ratio c_target / c_eddm is less than the waste ratio (P + D) / P. In plain language: if a targeted piece costs you twice an EDDM piece (premium ratio 2.0), targeting still wins as long as fewer than half the route is a real prospect (waste ratio above 2.0). On a mixed 30-year-old neighborhood where one home in three is due, the waste ratio is 3.0, so a targeted piece can cost up to three times an EDDM piece and still come out ahead on cost per job. That is why data-driven mail wins more often than the per-piece sticker shock suggests: the waste ratio in real neighborhoods is usually high. The two ways to lose the bet are a thin-prospect route where your data is wrong (you mail junk anyway) or a uniform route where nearly everyone is a prospect (waste ratio near 1.0, nothing to suppress).

Where the response numbers actually come from

One caution on the worked example above so you do not over-trust round numbers. The set rate (calls that become inspections) and close rate (inspections that become signed jobs) usually move more of your cost per job than the response rate does, and both are functions of who answered, not merely how many. A targeted list pulls a smaller raw call count but those callers are more often genuinely in-market, so the set and close rates on targeted calls tend to run higher than on a blanket drop where some callers are tire-kickers with a five-year-old roof. When you plug your own numbers in, do not assume the same set and close rates for both approaches. Pull them separately from your CRM if you have the history, because the quality difference in the calls is part of what you are buying with targeting.

How to build a roofing mail list that beats a blanket

Data-driven mail is only as good as the list. A lazy "all homeowners in ZIP 4XXXX" pull is barely better than EDDM and costs more. The edge comes from stacking filters that correlate with a roof that needs replacing soon. Here is the filter stack, roughly in order of impact.

Tier 1 filters: ownership and home type

Start by removing households that structurally cannot or will not buy a roof from you.

• Owner-occupied. For retail replacement, you want decision-makers who live in the home. Suppress renters; keep an absentee-owner segment only if you specifically pursue rental/landlord roofs with a different message.
• Single-family detached (and the small-multifamily you actually serve). Drop condos and HOAs that handle their own roofs unless that is your line of work.
• Property value band. Set a floor that matches your average ticket so you are not mailing homes that cannot finance a full replacement, and a ceiling if you are positioning as value rather than premium.

These three alone will often cut a route's mailable universe by a third to a half and remove most of the guaranteed non-responders EDDM would have charged you for.

Tier 2 filters: the roof's likely age

This is the filter EDDM cannot touch and where targeted mail earns its premium. A roof's life is the clock you are racing. The closer a roof is to the end of its service life, the more your mail piece reads like a timely reminder instead of an interruption.

There are three ways contractors approximate roof age, in increasing order of accuracy:

1. Home age as a proxy. Public assessor data gives a year built. On homes that never re-roofed, that is the roof age. The catch: a 1985 home may be on its second or third roof, so year-built overstates age and you mail people who just re-roofed.
2. Permit history. Some jurisdictions publish re-roof permits. Where available, a re-roof permit resets the clock and lets you suppress recently re-roofed homes. Coverage is patchy and pulling it is laborious.
3. Aerial-imagery roof-age estimation. Newer property-intelligence data estimates a roof's age as a range from current and historical aerial imagery, tile/shingle condition signals, and observed changes between image dates. This is the most direct signal because it reads the actual roof, not a proxy for it. It is a range and a probability, never a birth certificate, and you should treat it that way.

The honest framing on any roof-age data: it is a range, not a date. Good data will tell you a roof is likely 18 to 24 years old, not "installed on March 3." That range is exactly what you need for mail. You do not need the precise date to know a 20-to-25-year-old asphalt roof in a sun-and-hail climate is a live prospect. You need to separate the probably-due third of the neighborhood from the probably-fine two-thirds, and a range does that cleanly.

Tier 3 filters: storm exposure per property

Layer storm history on top of age and you get the sharpest roofing list available. Not "this ZIP got hail sometime," but which parcels sat under significant hail or damaging wind, how big the hail likely was, and how recently. Modeled storm exposure per roof lets you rank an entire territory by which roofs the weather most likely wore out, then mail from the top down until your budget runs out.

A few cautions so you use storm data like a pro and not a chaser:

• Storm exposure is odds, not proof. A high hail-exposure score means a roof was probably hit hard enough to matter, not that there is definitely a claimable loss. Your inspection decides damage; the data decides where to send the inspector.
• Hail swaths are narrow and patchy. Two streets apart can be the difference between 2-inch stones and nothing. ZIP-level storm mail wastes most of its postage; parcel-level storm targeting is the entire advantage.
• Recency matters for both relevance and, where insurance enters, for the homeowner's own filing timeline with their carrier. You document and estimate; the homeowner files and the carrier decides. (More on the compliance line below.)

Putting the stack together

The list you want is the intersection: owner-occupied single-family homes, in your value band, with a roof-age range in the back third of service life, ranked by storm exposure. That intersection is usually a small fraction of the EDDM universe, which is the point. You mail fewer, better houses, spend the saved postage on more touches, and your piece can honestly speak to the roof's situation because you selected for it.

A worked list-build, filter by filter

To make the funnel concrete, take a real-feeling territory and watch the universe shrink at each step. Start with a city ZIP that has 6,000 deliverable addresses, the number EDDM would charge you to blanket.

1. Owner-occupied single-family only. Drop renters, condos, and the small multifamily you do not service. Say that removes 40 percent. Universe: ~3,600.
2. Property-value band. Set a floor at the home value where a full replacement is realistically financeable for your average ticket, and a ceiling if you are positioned as value. Suppose that trims another 25 percent. Universe: ~2,700.
3. Roof-age range in the back third of service life. Keep only homes whose roof-age estimate puts them in, say, the 17-plus-year band for asphalt in your climate, and suppress any with a recent re-roof permit or a young aerial roof-age read. In a mixed-age city this is the big cut, often leaving 30 to 40 percent of what remained. Universe: ~1,000.
4. Storm-exposure rank. Score those ~1,000 by modeled hail and wind exposure and sort high to low. You do not necessarily cut here; you order. If your budget covers 600 pieces this month at three touches, you mail the top 600 by exposure and hold the rest for next cycle.

You went from 6,000 blanket addresses to a ranked 1,000, then mailed the top 600 three times for, very often, less total spend than a single 6,000-piece EDDM blanket. Same offer, but every dollar landed on a roof that is plausibly aging out and plausibly storm-worn, and you have a held-back list ready for the next drop. That ordered, intersected, repeatedly-mailed list is the practical opposite of EDDM, and it is what "data-driven" should mean in practice rather than as a buzzword.

Where the underlying data comes from

Know your sources so you can judge a vendor. Ownership, home type, value, and length of residence come from county assessor and recorder records, aggregated and refreshed by data resellers. Year-built also comes from the assessor. Re-roof permits come from municipal building departments, where published. Storm history at the regional level is public from the national weather service storm-events database and the storm prediction center, but turning regional reports into a per-parcel exposure score requires modeling hail size and wind over the parcel rather than reading a county-wide "hail reported" flag. Aerial roof-age estimates come from current and historical imagery analyzed for roof condition and change over time. None of these is perfect alone; the edge is in stacking them and, critically, in suppressing the recently-re-roofed homes that year-built data alone would wrongly keep.

This is the natural place a tool like RoofPredict fits. RoofPredict is built to tell roofing contractors which roofs are due, house by house: a roof-age range per address estimated from aerial imagery, plus storm physics modeled per roof so you can see which roofs the storm most likely wore out and which are simply aging out. It is not a lead-buying service and it does not hand you a homeowner who is ready to sign. What it does is rank doors, routes, and lists, and enrich your own CRM or mailing list with roof-age and storm signals, so your data-driven mail is pointed at the right third of the neighborhood instead of a blanket. The honest limits matter: the age is a range, the storm read is odds not a guarantee of damage, and your crew's inspection is still what confirms whether a roof actually needs work. Used that way, it turns the "is my selection good enough to beat EDDM" question from a guess into something you can see on a map before you spend a dollar.

When EDDM still wins

A fair piece has to say when the cheap blanket is the smart call, because sometimes it is.

• Uniform, high-density prospect routes. A subdivision where nearly every home is the same age and took the same storm has such high prospect density that suppression saves almost nothing. Blanket it.
• Brand-new shop, zero list infrastructure. If you have no CRM, no data vendor, and a storm just hit, EDDM lets you be in mailboxes tomorrow. Speed can beat precision in the first 72 hours after a storm. Build the data muscle afterward.
• Pure brand awareness. If the goal is name recognition before a busy season rather than this-week response, broad cheap reach has a role, though usually as a supplement, not the main play.
• Tiny mailable universe after filters. If your filters cut a route down to 30 homes, the printer's minimums and the per-piece data cost can make targeting uneconomic for that route. Sometimes the route is small and uniform enough that EDDM is just simpler.
• When your data is unreliable. Bad targeting is worse than no targeting because you pay the premium and miss. If you cannot trust your roof-age or storm source, do not pay for selection you cannot verify. Either fix the data or blanket.

The through-line: EDDM wins on speed, simplicity, and high prospect density. Data-driven mail wins on efficiency, repetition budget, and message relevance whenever the good prospects are a minority of the route and your selection is trustworthy.

Storm-season mail and the compliance line you cannot cross

Storm restoration is where roofing direct mail makes the most money and where contractors get themselves in the most legal trouble. The mail itself is fine. The language is where shops cross a line into territory regulated as public adjusting in many states, and that can carry real penalties. Whether you mail EDDM or a targeted storm list, the copy rules are identical, so learn them once.

Here is the clean mental model. A roofing contractor may:

• Inspect a roof and document what they find with photographs and measurements.
• Prepare an accurate, itemized repair estimate (commonly aligned to the Xactimate line-item structure carriers recognize) for the work they would perform.
• State facts about their own scope to a carrier when appropriate.
• Hand that documentation and estimate to the homeowner, who then decides what to do with it.

A roofing contractor may not, especially for a fee:

• Negotiate, adjust, or "handle" the homeowner's insurance claim.
• Interpret the homeowner's policy or tell them what is and is not covered.
• Promise a specific payout, an approval, or that a claim will be accepted.
• Promise that the deductible will be waived, absorbed, eaten, or otherwise made to disappear. Offering to cover or rebate a deductible to induce a claim is illegal in many states and is a fast way to lose a license and invite fraud charges.
• Advertise a "free roof" or imply the homeowner pays nothing.
• Represent the homeowner against their insurer. That is unlicensed public adjusting.

The safe frame for every storm mail piece is the same: we document thoroughly and write an accurate repair estimate; you file with your insurer and your insurer decides coverage. You are selling inspection and documentation quality, not a claim outcome.

A do-not-say list for your mail copy

Pin this near whoever writes your postcards. Every phrase on the left is a liability; the right column captures the same intent legally.

Notice that none of the legal versions are weak. "We will be on your roof tomorrow, document every hit with photos, and hand you an itemized estimate you can take to your insurer" is a stronger offer than "free roof," because it is specific, credible, and it survives a regulator reading your mailbox. Targeted storm mail makes this even better: when you mail only the parcels with real modeled hail exposure, your documentation promise is believable because you are knocking on roofs that probably took a hit.

The documentation workflow you are actually selling

Since the legal offer is documentation and estimate quality, it helps to spell out the workflow your mail is promising, because doing it well is what earns the referrals that eventually shrink your mail spend. A tight storm-inspection-and-estimate process looks like this:

1. Set the inspection from the call. Capture the address, confirm it is owner-occupied, and note the storm date you are referencing. If you mailed from a storm-exposure list, you already know the parcel modeled real impact, which makes the inspection worth the crew's time.
2. Document the whole roof, not only the obvious slope. Photograph each elevation, the field, the ridges, the flashings, the vents, and the soft metals (gutters, downspouts, screens, AC fins) where hail leaves clear directional bruising. Soft-metal evidence is often the clearest signal that hail of a damaging size fell, and it is fact, not opinion.
3. Mark and measure. Chalk test squares, count impacts per square where you find them, and record slope and stories. Pull roof measurements (from a measurement report or careful field measure) so your estimate quantities are defensible.
4. Write the estimate to the line-item structure carriers recognize. Itemize tear-off, underlayment, the shingle and accessories, flashing, ventilation, and the labor and disposal lines, in the Xactimate-aligned format adjusters work in. You are estimating your repair scope as facts, not interpreting the policy.
5. Hand the package to the homeowner. Photos, measurements, and the itemized estimate go to the homeowner. They decide whether to file. If they file, the insurer assigns an adjuster and the insurer decides coverage. Your role is the documentation and the repair, not the claim.

Every step there is something a contractor may legally do and do well. None of it requires you to interpret coverage, promise an outcome, or touch the deductible. When your mail promises this workflow and your crew actually delivers it, the homeowner gets a credible package and you get a clean, lawful position, which is exactly the reputation that turns one job into three on the same street.

Why parcel-level storm targeting beats blanket storm EDDM on compliance, too

There is a quieter benefit to parcel-level storm data beyond efficiency. When you blanket a whole town with "storm damage" mail after a regional hail report, a large share of those roofs were never actually hit, and your piece starts to feel like the predatory storm-chaser mail that prompts state regulation in the first place. When you mail only the roofs that modeled high exposure, your message is honest by construction: you are contacting people whose roofs probably took real hail, offering documentation, not a payout. Tighter data keeps you on the right side of both economics and ethics.

Designing the piece: what to put on the postcard

Targeting decides who; the piece decides whether they call. A few rules that hold across EDDM and data-driven roofing mail.

• One offer, one action. A free documented inspection, or a roof-age check, or a neighborhood-proof callback. Not all three crammed on one card.
• Make the relevance obvious in three seconds. With data-driven mail you can reference the roof's likely situation ("roofs in your neighborhood built in the late 1990s are reaching the age where we usually find wear") in a way EDDM cannot, because EDDM has no idea who is reading it.
• Local proof over national polish. A real local phone number, a recognizable nearby street, a license number, and a photo of an actual local job beat slick stock photography.
• A reason to act now that is honest. Roof age ("asphalt roofs in this climate typically run 18 to 25 years") or a recent verified storm. Avoid manufactured urgency and avoid any claim-outcome promise.
• Big, mobile-friendly response paths. Phone number large, plus a QR code or short URL, plus a text option if you answer texts. Make the easy action the obvious one.
• Size and format that survive the mail. The large EDDM-flat format stands out in a stack; verify your printer's piece meets current USPS dimensional rules for whichever rate you are using.

Sequencing: the second and third touch is where the money is

The single biggest mistake in roofing mail is the one-and-done drop. Response compounds with repetition, and the budget you save by targeting is exactly what funds repetition. A simple, effective cadence for a data-driven list:

1. Touch 1 (week 0): Introduction and offer. Roof-age or storm relevance line. Free documented inspection.
2. Touch 2 (week 2 to 3): Local proof. "We just completed roofs near you" with a real street or photo. Same offer, new angle.
3. Touch 3 (week 4 to 6): Light urgency tied to season or the roof's age range. Last-call framing without false scarcity.

Three honest touches to a sharp list of likely-due homes will, in most markets, out-convert a single EDDM blanket to ten times as many houses, for comparable money. That is the entire strategic case for data-driven mail in one sentence: fewer houses, more often, chosen because their roofs are actually due.

Pairing mail with the door and the phone

Mail rarely works hardest alone. The shops that win treat the mailing list as the spine of a small multi-channel sequence. Because a data-driven list is a known set of addresses, you can hand the same ranked list to a canvassing crew and have them knock the streets you just mailed, referencing the piece ("you may have gotten our card") to warm the door. You can append phone numbers to part of the list for a follow-up call on the highest-exposure homes. You can run a tight geo-fenced ad to the same neighborhoods so the brand looks familiar when the postcard lands. EDDM supports none of this cleanly because you never knew which specific households you reached. This is a quiet but real advantage of the data-driven path: the list is reusable across channels, and each channel makes the others convert better. Knock the door of a home that got three of your postcards and modeled high hail exposure, and you are having a very different conversation than a cold first knock.

A decision framework you can run this week

Stop debating EDDM vs. data-driven in the abstract and decide per route. Walk this checklist for each neighborhood you are considering.

Step 1 — Estimate prospect density. What share of homes on this route plausibly need a roof this year? Use home-age data, your knowledge of the area, and storm history. If it is high (say more than half, like a uniform post-storm subdivision), lean EDDM. If it is low to moderate, lean data-driven.

Step 2 — Check your data quality. Do you have a roof-age range and/or parcel storm exposure you trust for this area? If yes, targeting is viable. If no, either get the data or default to EDDM rather than paying for a guess.

Step 3 — Run both cost-per-job models. Use your real per-piece costs and honest set/close rates. Compute total spend and cost per job for an EDDM blanket vs. a targeted subset. Let the number decide, not the habit.

Step 4 — Budget for repetition. Whichever you pick, plan at least two to three touches. If targeting frees up budget, spend it on more touches to the good list, not on widening to weaker houses.

Step 5 — Lock the copy to the compliance line. If storm or insurance language appears anywhere, run it against the do-not-say table. Sell documentation and estimate quality, never a claim outcome, a waived deductible, or a free roof.

Step 6 — Track at the job level. Use a dedicated phone number and/or QR per campaign so you can attribute calls. Measure cost per inspection and cost per signed job, by route, by approach. Feed that back into Step 1 next season.

A simple tracking sheet that pays for itself

You do not need fancy attribution software to run this well. A spreadsheet with one row per campaign and these columns will outperform most shops' guessing: campaign name, route or list ID, approach (EDDM or targeted), pieces mailed, touches, total spend, calls received, inspections set, jobs signed, total contract value. From those you compute the four numbers that matter: cost per call, cost per inspection, cost per signed job, and return on mail spend (contract value divided by spend). Sort your finished campaigns by cost per signed job and you will see your own pattern emerge, which routes reward blanketing and which reward targeting, faster than any article can tell you. The shops that keep this sheet for two seasons stop arguing about EDDM versus data-driven entirely, because they can read the answer off their own history per neighborhood.

Seasonality and timing the calendar

One more variable the framework should respect: the calendar. Roof-replacement demand and storm activity both swing seasonally, and your mail should lead the demand, not chase it. Aging-out retail replacement mail works best landing a few weeks ahead of your busy installation season, when homeowners are starting to think about the house but before they are calling everyone in town. Storm mail is the opposite, it is reactive and time-sensitive, landing in the days-to-weeks window after a verified event when the problem is fresh and the homeowner's own filing timeline with their carrier is still open. A practical pattern many shops run: data-driven aging-out mail as the steady base load through the season, with the ability to flip to fast post-storm mail (EDDM for a uniform hard-hit subdivision, parcel-targeted for a patchy swath) when the weather creates the opening. Keep both motions ready and let the sky decide which one fires.

A quick reference table

What pros get wrong

Even shops that mail a lot make the same handful of mistakes. Avoiding these matters more than picking EDDM vs. data-driven.

• Judging by cost per piece. Already covered, but it is the cardinal error. The cheap piece that reaches the wrong people is the expensive campaign. Always reason in cost per signed job.
• One drop and done. Single touches underperform sequences badly. If your budget only allows one drop to a big list, you would usually do better with three drops to a third of the list.
• ZIP-level storm targeting. Hail is a swath, not a ZIP. Blanketing a ZIP after a regional hail report wastes most of the postage on roofs that were never hit and erodes your credibility. Parcel-level or it is barely better than guessing.
• Year-built as gospel. Treating year-built as roof age over-mails homes that recently re-roofed. Suppress known recent re-roofs where you can, and prefer aerial-based roof-age ranges that read the actual roof.
• Confusing a list with leads. A data-driven list, even a great one, is a ranked set of doors, not a stack of ready buyers. The list tells you where to spend attention. Your offer, your follow-up, and your inspection close the job. Any vendor promising "leads" from a mailing list is overselling.
• Compliance as an afterthought. Writing copy first and checking legality later is how "free roof" and "we waive the deductible" end up in print. Bake the do-not-say list into your template so the dangerous phrases never get drafted.
• No tracking. Mailing without a campaign-specific phone number or code means you are guessing at what worked. You cannot improve Step 1 next season if you never measured this one.

Bringing it together

Every Door Direct Mail and data-driven mail are not rivals to crown a winner between. They are two settings on the same dial, and the dial is how much of the route is actually a prospect. When nearly everyone on the route has a roof that is due or storm-worn, blanket it cheaply with EDDM and move fast. When the roofs that matter are a minority, scattered across mixed-age streets or under a narrow hail swath, pay the premium to select them, speak to their actual situation, and reinvest the saved postage into the second and third touches that win the job.

The lever that has gotten dramatically better in recent years is the selection itself. Estimating a roof-age range per address from aerial imagery and modeling storm exposure per parcel turns "which third of this neighborhood is due" from a hunch into a map. That is exactly the gap where RoofPredict is meant to help: it ranks doors, routes, and lists by which roofs are aging out and which the storm most likely wore out, and enriches your own CRM or mailing list with roof-age and storm signals so your data-driven mail lands on the right houses. It will not hand you a signed contract, the roof age is a range and the storm read is odds rather than proof, and your crew's inspection still decides the truth on each roof. Used honestly, with copy that documents and estimates rather than promising claim outcomes, it makes the cost-per-job math tilt your way far more often than a blanket ever could.

Pick the tool per route, run the real numbers, mail the good houses more than once, and keep every storm-season word on the safe side of the line. That is the whole game.

FAQ

Is EDDM or data-driven mail cheaper for roofing?

EDDM is almost always cheaper per piece because it skips the list cost and uses a lower retail postage rate. But cheaper per piece is not cheaper per signed job. If your targeting accurately identifies the roofs that are actually due, data-driven mail often costs less per acquired job even though each piece costs more, because you stop paying to reach homes that will never call. Run both cost-per-job models with your real per-piece costs and set/close rates before deciding.

What response rate should I expect from a roofing postcard?

There is no single number. Industry-wide direct mail response to cold prospect lists generally runs well under 1%, with house lists (people you have touched before) responding higher. Roofing response then multiplies with roof-age relevance, post-storm timing, local proof, and repetition. The reliable way to lift response is to improve who gets the mail and when, not to chase a clever headline. Track cost per inspection and per signed job rather than fixating on a percentage.

Can I target direct mail by roof age?

Yes, with caveats. You can approximate roof age three ways: year-built from assessor data (a proxy that overstates age on homes that re-roofed), re-roof permit history where a jurisdiction publishes it, or aerial-imagery roof-age estimation that reads the actual roof and returns an age range. Aerial-based estimates are the most direct because they look at the roof itself, but the output is always a range and a probability, not an exact installation date. A range is enough to separate the likely-due homes from the rest.

How accurate is aerial roof-age data?

It is a calibrated estimate, not a record. Good aerial-imagery roof-age data gives you a range, for example likely 18 to 24 years old, derived from current and historical imagery and condition signals. Treat it as a way to rank and segment a mailing list, not as proof of a roof's exact age. Your crew's on-roof inspection is still what confirms the real condition. The value is in cleanly separating the probably-due third of a neighborhood from the probably-fine remainder before you spend on mail.

What can I legally say about insurance in roofing mail?

You can offer to inspect, document damage with photos, and prepare an accurate itemized repair estimate the homeowner can share with their insurer. You cannot promise a payout or approval, say a loss is covered, promise to handle or negotiate the claim, advertise a free roof, or promise to waive or absorb the deductible. Those cross into unlicensed public adjusting or illegal inducement in many states. The safe frame: you document and estimate; the homeowner files and the insurer decides coverage.

Is it illegal to offer to waive the deductible?

In many states, yes, offering to waive, absorb, rebate, or otherwise eliminate a homeowner's insurance deductible to induce a claim is illegal and can carry license loss and fraud exposure. The deductible is the homeowner's responsibility. Keep any mention of the deductible out of your mail entirely. Sell the quality of your inspection and documentation instead, which is a stronger and fully legal offer.

Why is parcel-level storm data better than mailing a whole ZIP after a storm?

Hail falls in narrow, patchy swaths; two streets apart can be the difference between damaging stones and nothing. Mailing a whole ZIP after a regional hail report wastes most of your postage on roofs that were never hit and makes your storm-damage message feel predatory. Parcel-level storm exposure lets you mail only the roofs that modeled real impact, which is both more efficient and more honest, because you are contacting people whose roofs probably took a hit and offering documentation rather than a payout promise.

Does a data-driven mailing list give me leads?

No. A list, even an excellent one, is a ranked set of doors, not a stack of ready buyers. It tells you which houses are most likely to have a roof that is due or storm-worn, so you spend attention where it pays off. Your offer, your follow-up sequence, and your inspection are what turn a door into a job. Any vendor selling a mailing list as guaranteed leads is overselling what list data can do.

How many times should I mail the same household?

Plan for at least two to three touches. Single drops underperform sequences badly because response compounds with repetition. A practical cadence: an intro-and-offer touch, a local-proof touch a couple of weeks later, then a light season-or-age urgency touch a few weeks after that. The budget you save by targeting a smaller, sharper list is exactly what should fund these extra touches to the good homes rather than widening the mailing to weaker ones.

When does EDDM still make more sense than targeting?

EDDM wins when prospect density is high and uniform, like a subdivision where nearly every home is the same age and took the same storm, so suppression saves almost nothing. It also wins when you are a brand-new shop with no list infrastructure and a storm just hit and speed matters most, when a route filters down to too few homes to be economical, or when your roof-age and storm data are not trustworthy. Bad targeting is worse than no targeting because you pay the premium and still miss.