Is Roof Targeting Data Worth It? An Honest ROI Breakdown for Contractors

Every roofing owner I know has been pitched some version of "buy our data and stop wasting money knocking the wrong doors." The pitch is always clean. The math is almost never shown. So you sign up, you spend three months feeding a list to your canvassers or stuffing it into a mailer, and you're left squinting at your QuickBooks trying to figure out whether the thing actually paid for itself or just felt productive.

I've run that experiment more than once, on more than one platform, in markets with storms and markets without. The honest answer to "is roof targeting data worth it" is the same answer you'd give about a new truck or a new salesperson: it depends entirely on what you're replacing, how disciplined you are about measuring, and whether your sales motion can actually convert a better list. The data itself doesn't close anybody. It changes the odds at the top of your funnel, and a small change at the top compounds — or evaporates — depending on everything downstream.

What follows is the breakdown I wish someone had handed me before my first subscription: the real cost components beyond the line item, the value levers that actually move money, a break-even framework you can run on a napkin, the situations where it does NOT pay off, and how to measure payback without lying to yourself. I'll use one platform, RoofPredict, as a worked example in a couple of spots because I know its model well, but the framework is tool-agnostic and applies whether you're looking at property data, roof-age estimates, storm overlays, or a blended canvassing list.

What "roof targeting data" actually means (and what it doesn't)

Before you can price the ROI, you have to be precise about what you're buying, because the category is muddy and vendors lean into the mud.

Roof targeting data, broadly, is property-level information that helps you decide which house to work next instead of working the whole street. It usually blends some mix of:

• Roof age or roof-age band — derived from permit records, aerial imagery change-detection, or year-built as a (weak) proxy. The honest versions give you a range ("15–20 years"), not an exact install date, because nobody has a clean national database of every re-roof.
• Storm exposure — hail and wind history mapped to the property, sometimes as a simple "a storm passed through this ZIP" overlay, sometimes as a per-roof impact model. There's a real difference between those two, and it matters for ROI.
• Property attributes — square footage, stories, ownership status, owner-occupied vs. rental, sometimes a rough home value or equity signal.
• A ranking or score — the vendor's attempt to stack-rank homes by how "due" or "worth your time" they are.

Here's what it is NOT, and confusing these is the single most common reason people feel burned:

• It is not a lead service. Targeting data tells you which doors to knock or which addresses to mail. It does not hand you a homeowner who raised their hand. If you expected inbound and bought a list, you'll conclude (wrongly) that "the data doesn't work."
• It is not a measurement product. EagleView, HOVER, Roofr, and similar tools measure a roof you've already picked. Targeting data picks the roof. Different category, different job. People conflate them and then can't figure out why their EagleView reports didn't grow the pipeline — measurement was never supposed to.
• It is not condition truth. Year-built from a county assessor (or from Zillow/Google) tells you when the house was built, which says nothing about a re-roof done eight years ago. A storm forecast is odds of damage, not proof of damage. Roof age from imagery is an estimate. Treat all of it as a probability improvement, not a guarantee.

Getting this straight is itself part of the ROI question, because the value of the data is only ever "how much does it improve the odds versus what I'm doing now." If your current baseline is already good, the lift — and the ROI — is smaller. If your baseline is "we knock the whole subdivision," the lift can be large.

The cost components nobody puts on the invoice

The subscription price is the part everyone fixates on and it's usually the smallest real cost. If you only budget the line item, your ROI math will be wrong in the optimistic direction. Here are the components that actually make up your total cost of ownership.

1. The subscription / data fee

This is the obvious one. Pricing in this category ranges widely depending on whether you're buying a flat monthly platform fee, a per-record or per-report charge, per-market licensing, or some bundle that includes mail or a CRM. I'm not going to quote specific numbers as fact because they move and they vary by vendor and market — get a real quote for your area. The thing to nail down before signing:

• Is it per record, per market, or flat? Per-record looks cheap until you pull a big list.
• Are there refresh costs? Roof age and storm data go stale. A one-time pull degrades.
• What's the contract term and out clause? Annual lock-ins change the break-even math because you're committing the spend whether or not month one works.

2. The labor to act on it

This is the cost that sinks naive ROI estimates. A list of 2,000 "due" roofs is worthless until somebody knocks them, mails them, or calls them. That's payroll, gas, and management attention. If a rep can productively work, say, a few dozen quality doors a day, a 2,000-home list is weeks of field labor. The data didn't reduce the work of doing the work — it (hopefully) made each unit of work more likely to land.

Budget the fully-loaded cost of the motion you'll run against the list: canvasser hourly + payroll tax + vehicle, or the all-in cost of a mail drop (more on that below), or the SDR time to dial.

3. The learning curve and integration tax

There's always a ramp. Your team has to learn to trust the list, learn the app or the export format, and you have to wire it into how you already operate — your CRM, your routing, your mail vendor. If the data lives in a portal nobody opens, you paid for a subscription you don't use. Realistically budget 30–90 days where you're paying full price for partial adoption. I've watched shops cancel in month two and blame the data when the truth was they never finished onboarding.

If the tool offers two-way CRM sync (HubSpot, ServiceTitan, JobNimbus, AccuLynx, Jobber, and the like), that integration tax drops a lot, because the list shows up where your reps already live instead of as a spreadsheet that rots. That's a real, if unglamorous, ROI lever — adoption is most of the battle.

4. Opportunity cost and list fatigue

If your team spends 20 hours a week working a data list, that's 20 hours not spent on referrals, repeat customers, or your existing pipeline. Sometimes the data list is the better use; sometimes it isn't. And lists fatigue — knock the same "due" homes three times with no new angle and your contact rate craters. The cost of a list isn't just buying it; it's the management discipline to rotate and refresh it.

Total cost of ownership, summarized

Your real cost is the sum of all five, not the first row. Now let's look at how it pays you back.

A quick gut-check on cost before you go further

A rule of thumb I use to sanity-check any new top-of-funnel spend: the all-in cost of the program, divided by the number of jobs I realistically expect it to influence in a quarter, should land at or below my current cost per acquisition from my best existing channel. If buying data pushes my blended cost per job above what referrals and repeat work already cost me, the burden of proof is on the data to show a lift big enough to justify the premium. It's allowed to cost more per job than referrals — referrals don't scale on demand and data does — but you should know you're paying that premium on purpose, not by accident. Most owners who feel burned never did this single division, so they never noticed the channel was structurally more expensive than the one they already had.

The four value levers (where the money actually comes from)

Targeting data can pay off through four distinct mechanisms. They're not equal, they don't all apply to every shop, and you should know which one you're betting on before you buy. If you can't name your lever, you're guessing.

Lever 1: CAC reduction — fewer dollars chasing each won job

The core promise. If 30% of the homes on a generic street have a roof too new to need you, then knocking or mailing all of them means roughly a third of your spend is structurally wasted before anyone says a word. A list that filters out the obviously-not-due homes raises the density of viable prospects per hour of labor or per piece of mail. Same effort, more shots at a real opportunity, lower cost per acquisition.

The size of this lever depends entirely on your baseline. If you already work tight, hand-picked neighborhoods you know well, the data's filtering adds little. If you carpet-bomb, it can add a lot.

Lever 2: Close-rate lift — better conversations, beyond better lists

This one is underrated and it's where the per-home report features earn their keep. When a rep knocks with a specific, credible reason — "your roof's in the range where we see these failing, and there were two hail events on this block in the last few years" — the conversation is warmer than "hi, we're doing roofs in the neighborhood." A homeowner-facing report, a microsite, a PDF with a QR code: these turn a cold knock into something closer to a consultation.

The second-order effect here is rep retention, which quietly drives ROI more than people credit. A green canvasser armed with a per-home talking point sounds competent on day three instead of week twelve. They get a win sooner, they make money sooner, and they stay. Replacing a canvasser costs you recruiting time, ramp, and lost production. Anything that helps a new hire close earlier is worth real money even though it never shows up as a "data" line item.

Lever 3: Fewer wasted truck rolls and inspections

Every roof you climb that turns out to have five years of life left is a truck roll, an hour of a closer's time, and a no-sale. Better pre-qualification at the top means a higher share of your inspections are on roofs that are genuinely candidates. The labor savings here are concrete and easy to underweight because they're diffuse — a few avoided pointless inspections a week adds up over a season.

Lever 4: Recovered revenue from your own book and your own scope

Two flavors here, both real:

• Mining your existing CRM. The cheapest "new" customer is an old estimate that never closed or a past customer whose roof has now aged into the danger zone. Targeting data layered onto your own database re-surfaces homes you already paid to acquire once. This is some of the highest-ROI activity in the whole category because the acquisition cost is near zero — you're just reactivating.
• Tighter documentation on the work you already do. On the storm/restoration side, the value isn't in the targeting list at all — it's in not leaving money on the table during a claim you're already working. Tools that help a contractor OCR and organize its own inspection documents, run a photo and measurement checklist, score packet completeness, and compare its own estimate to the carrier's estimate internally can surface scope you documented but didn't capture. Important honesty and legal note: a contractor may document its own inspection, estimate, and evidence and request missing documents, but it may not represent, interpret, or negotiate the homeowner's claim, advise on coverage or settlement, or use deductible-waiver messaging — that's the public-adjuster line and it's licensed in most states. The homeowner files, the insurer decides, the contractor documents its own scope. Within that boundary, recovered supplement revenue on jobs you're already doing can be a meaningful, often-overlooked payback channel. (See your state DOI and the NAIC for where that line sits.)

Notice that Levers 1–3 are about getting more or cheaper jobs and Lever 4 is about capturing more from jobs you already have. Many shops over-focus on Lever 1 (the shiny "target better" promise) and ignore Lever 4, which is often the faster payback because it doesn't require any new field labor.

Which lever is yours? A quick self-diagnosis

Before you model any numbers, figure out which lever the data has to pull for your shop, because that determines whether it can pay and how fast. Run yourself through this:

• If you carpet-bomb whole neighborhoods today, your lever is CAC reduction (Lever 1). The data earns its keep by cutting the not-due homes out of your effort. Big potential lift, and the easiest to feel.
• If your reps are green or you have high canvasser turnover, your lever is close-rate and retention (Lever 2). The per-home report and talking point are what you're really buying. The payback shows up partly as faster ramp and lower rehiring cost, which is real money that's easy to miss because it never appears as a sales line.
• If your closers waste time on roofs with years of life left, your lever is fewer truck rolls (Lever 3). Watch your inspection-to-sale ratio; that's where this shows up.
• If you have a fat CRM full of old estimates and past customers, your lever is reactivation (Lever 4). Often the fastest payback in the whole category, because the acquisition cost is near zero and the labor is a phone-and-mail re-touch, not fresh canvassing.
• If you do storm restoration work, you likely have a Lever 4 documentation angle on jobs you're already running — recovered scope within the contractor-documentation boundary.

Most shops have two or three of these in play at once. The mistake is buying for the loud lever (1) and never bothering to pull the quiet ones (4), then concluding the data was a wash when really you only used a quarter of what you paid for.

A break-even framework you can run on a napkin

Here's the part that actually answers the question. You don't need the vendor's case study; you need your own numbers. Below is the framework I use. Every number in the worked example is an illustrative assumption I'm stating openly so you can see the mechanics — do not treat them as measured facts or as typical results. Plug in your real figures.

Step 1: Know your average job profit, not revenue

Use gross profit per won job, not the ticket price. If your average residential re-roof brings, say, $4,000 in gross profit after materials and labor (your number will differ — pull it from your books), that $4,000 is what each incremental win is worth to the ROI calculation. Using revenue here is the most common way people fool themselves into thinking a channel pays.

Step 2: Compute the total cost of the program

Add the five cost components from earlier for the period you're measuring (say, one quarter):

Total program cost = data subscription
                   + fully-loaded labor to work the list
                   + onboarding/ramp time (amortized)
                   + data refresh
                   + (optional) mail/print/postage if you're mailing

Step 3: Find your break-even job count

Break-even wins = Total program cost / gross profit per won job

Worked example (illustrative assumptions, not facts): Suppose for one quarter your all-in program cost — subscription plus the canvasser hours plus a modest mail drop — comes to $12,000. At $4,000 gross profit per job, you need 3 incremental wins to break even.

The word that's carrying all the weight is incremental. Which brings us to the step everyone skips.

Step 4: Isolate INCREMENTAL wins (the honesty step)

The wins that count are only the ones you would NOT have gotten anyway. If your canvassers would have knocked somewhere this quarter and closed some jobs regardless, you can't credit all of those to the data. You have to estimate the lift over baseline.

Two defensible ways to do it:

1. Hold-out test. Run a control. Have one crew work data-targeted homes and a comparable crew work your normal method in a similar area for the same period. The difference in close rate or jobs-per-rep-hour is your true lift. This is the gold standard and almost nobody does it. Do it.
2. Before/after on a stable metric. If a clean hold-out isn't practical, track a ratio that isolates efficiency, like jobs won per 100 doors knocked or cost per acquisition, before and after the data, holding everything else constant. A move from, say, 2 wins per 100 knocks to 3 is a 50% efficiency lift — and that delta is what the data bought you, not your total volume.

So the honest ROI is:

Program ROI = (Incremental wins x gross profit per job - Total program cost)
              / Total program cost

Worked example continued (illustrative): Say over the quarter the targeted crew won 9 jobs and your hold-out math suggests 4 of them were incremental (the other 5 you'd plausibly have gotten anyway). 4 incremental wins x $4,000 = $16,000 of incremental gross profit against $12,000 of cost. That's a positive ROI of about 33% for the quarter, and a payback that clears in the quarter. Change the assumption to only 2 incremental wins and you're underwater — same data, same cost, different lift. The entire decision lives in that incrementality number, which is exactly why you have to measure it instead of vibing it.

Step 5: Mail has its own break-even (run it separately)

If your motion is direct mail, do a parallel calculation, because mail ROI is brutally sensitive to response rate. Tracked direct mail typically converts at a low single-digit-percent response or less, and the targeting is supposed to push that response rate up by removing the not-due homes from the drop. Your mail break-even:

Mail break-even response rate = Cost per piece
                              / (gross profit per job x close rate on responders)

If each mailed piece costs you a dollar-something all-in (design, print, postage, list), and a responder closes at some rate into a $4,000-profit job, you can solve for the response rate you need to break even, then ask honestly whether better targeting can plausibly get you there. The USPS publishes postage and Every Door Direct Mail specifics if you're modeling mail costs precisely. The point: a smaller, better-targeted drop with a higher response rate often beats a big cheap blanket drop, but only if the targeting lift is real — measure it with tracked mail and per-piece delivery confirmation, not by feel.

Step 6: Sensitivity-test the two assumptions that swing the answer

A break-even number is only as trustworthy as the two inputs that move it most: gross profit per job and incremental lift. Before you commit, run the calculation three times — pessimistic, expected, and optimistic — on those two inputs and see whether the decision flips.

(Every figure here is an illustrative assumption, not a benchmark.) If the program only pencils out in the optimistic column, you're making a bet, not a sound investment — proceed cautiously and with a short leash. If it's positive even in the pessimistic column, it's a comfortable buy. Most real decisions sit in the middle, where the verdict hinges on whether your true incremental lift lands closer to 2 or 4 — which is exactly why Step 4 isn't optional. The owners who get burned almost always ran a single optimistic line and called it the plan.

A note on what counts as "a win"

Decide up front whether a win means a signed contract, a completed installation, or collected cash, and use the same definition everywhere in the math. Signed contracts overstate the picture because some cancel; collected cash is the truest but lags by weeks. I default to signed contracts net of your historical cancellation rate — it's timely enough to make decisions on and honest enough not to flatter the result. Whatever you pick, freeze it for the whole evaluation so you're not comparing signed-contract wins this quarter against collected-cash wins last quarter and fooling yourself with a moving definition.

When roof targeting data does NOT pay off

This is the section vendors skip and the one that'll save you the most money. Targeting data is a bad buy in several specific situations. If any of these is you, keep your wallet closed until it changes.

• Your sales motion can't convert. Data sharpens the top of the funnel. If your closers are weak, your follow-up is sloppy, or you have no consistent canvassing or mail discipline, a better list just gives you more well-qualified prospects to mishandle. Fix the motion first. The highest-ROI move for a shop with a leaky funnel is rarely more data.
• You're too small to feed it. If you can't field the labor to work a list of even a few hundred homes, you'll pay for data you can't act on. Below a certain volume, your own referrals and repeat customers are a better use of every hour. Data scales effort you already have; it doesn't create effort.
• Your baseline is already tight. Veteran owners who hand-pick neighborhoods from decades of local knowledge sometimes already have an internal "targeting model" in their heads. The data's lift over a genuinely good baseline can be too small to clear its cost. Be honest about how good your baseline really is — most people overrate it, but a few don't.
• You bought it expecting leads. Worth repeating because it's the #1 cause of "it didn't work." Targeting data is fuel for outbound. If you wanted inbound, you bought the wrong category and no amount of data quality will fix the mismatch.
• You won't measure. If you're not willing to run a hold-out or at least track a clean efficiency ratio, you will never actually know if it paid, you'll renew or cancel on gut feel, and you'll have wasted the spend either way. Unmeasured data spend is just a different flavor of the blanket-marketing waste you were trying to escape.
• The data is stale or wrong for your area. Targeting quality varies by market. Permit-record coverage is uneven across counties, storm modeling is better in some regions than others, and roof-age estimates from imagery are only as good as the imagery refresh. Always run a sample on streets you know cold before you trust a list on streets you don't. If the vendor won't let you sanity-check a sample against your own knowledge, that's a tell.

How to measure payback without lying to yourself

Assume you buy it. Here's the measurement discipline that separates owners who actually know their ROI from owners who have a feeling about it.

Set the baseline before you start

You cannot measure lift if you don't know your starting point. Before the data arrives, write down, for your current method: jobs won per rep per week, doors knocked or pieces mailed per win, your blended cost per acquisition, and your inspection-to-sale ratio. If you don't have these, spend two weeks measuring your status quo first. A baseline captured after you've already changed things is worthless.

Use the right denominator

Cost per acquisition is the cleanest single metric for this decision because it folds cost and outcome into one number. But compute it consistently — same definition of "cost," same definition of "acquisition" (signed contract? completed job? collected payment?) before and after. Quietly changing the denominator is the most common self-deception in marketing ROI.

Track the full funnel, not only wins

A single number hides where the value (or the leak) is. Watch the whole chain:

1. Contact rate — are you reaching a person? (Mostly a function of your motion, not the data.)
2. Conversation-to-inspection rate — is the targeting producing warmer, more qualified conversations? (Here the data should help.)
3. Inspection-to-sale rate — are the roofs you climb actually candidates? (Strong signal the targeting is working.)
4. Cost per win — the bottom line.

If cost per win improved but inspection-to-sale didn't, your gain probably came from somewhere else and the data may not deserve the credit. The funnel keeps you honest about attribution.

Give it a real, time-boxed trial

Don't evaluate at 30 days — you're still inside the onboarding ramp and the numbers are noise. Don't drift for a year with no decision either. Commit to a 90-day trial with a pre-agreed success threshold ("cost per win drops below X" or "we clear N incremental jobs"), then make a clean keep/kill call. Writing the threshold down before you start removes the temptation to rationalize a renewal after.

Separate the levers in your reporting

If you're using the data for both new outbound (Levers 1–3) and CRM reactivation (Lever 4) and supplement documentation on storm jobs, tag the resulting wins by source. Reactivation and supplement recovery often pay back faster and look very different from cold outbound. Blending them hides which part is actually carrying the program and leads you to keep or kill the whole thing when you should be doubling down on one lever and dropping another.

Common attribution traps that inflate the data's credit

Four ways the numbers lie to you, all of which I've fallen for at least once:

1. The seasonality trap. If you turn on the data in spring as your busy season ramps, your wins go up — but they'd have gone up anyway. Compare like seasons, or use a same-period hold-out, so you don't credit the data for the calendar.
2. The storm trap. A hail event lands mid-trial and your phone rings off the hook. Those wins are the storm's, not the targeting list's. Tag and exclude storm-driven inbound from your cold-targeting lever math, or you'll renew on a number that won't repeat in a quiet quarter.
3. The best-rep trap. You hand the shiny new list to your strongest closer because you want it to succeed. Now you can't tell the lift from the rep. Assign the data to an average crew, or rotate it, if you want a number you can trust.
4. The double-count trap. A homeowner from your CRM reactivation also happens to be on the cold-targeting list, and both levers claim the win. Pick a single attribution rule (first-touch is simplest) and apply it consistently so your levers sum to your actual total instead of overstating it.

None of this requires fancy software — a tagged column in your CRM and the discipline to fill it in honestly covers it. The tools that already track cost-per-win and actual-versus-estimate in a results view save you the spreadsheet, but the discipline matters more than the tooling.

Build a one-page scorecard you actually look at

Make the measurement boringly repeatable or it won't happen. One page, updated weekly, with these rows: doors worked or pieces mailed, contact rate, conversation-to-inspection rate, inspection-to-sale rate, jobs won (tagged by lever), program cost to date, and cost per win. The first time most owners see all of these next to each other, they spot the leak immediately — usually it's not the data, it's a conversion step downstream that the data can't fix. That diagnosis alone is often worth more than the subscription.

A worked example, end to end

Let me put the whole framework together on one illustrative shop so you can see the shape of it. Every figure here is an openly-stated assumption for demonstration, not a benchmark or a promise.

The shop: mid-size residential roofer, two canvassing crews, a CRM with a few thousand old estimates, working a mix of storm and non-storm neighborhoods. Average gross profit per won job: $4,000.

The program (one quarter):

• Data subscription: $3,000
• Two canvassers, partial allocation to the targeted list: $6,000 fully loaded
• A tracked mail drop to the top-ranked non-storm homes: $2,500 all-in
• Onboarding/ramp, amortized: $500
• Total program cost: $12,000

Break-even: $12,000 / $4,000 = 3 incremental wins to break even.

What they measured (with a rough hold-out):

• Targeted crew: inspection-to-sale rate rose from 25% to 33% versus the control crew working their normal method — a real, isolated lift attributable to better pre-qualification.
• Net incremental cold-outbound wins over baseline: 3.
• CRM reactivation (Lever 4): the data flagged 40 past estimates whose roofs had aged into range; a phone-and-mail re-touch closed 2 of them — near-zero acquisition cost, so essentially pure margin.
• Supplement documentation on storm jobs they were already doing: tighter internal scope-to-carrier comparison and packet completeness surfaced documented scope on a couple of jobs worth, say, a few thousand dollars of recovered gross profit (within the contractor-documentation boundary above).

The honest ROI: the cold-outbound lever roughly broke even on its own (3 wins = break-even). The program turned clearly positive because of Levers 4 — the CRM reactivation and the supplement documentation — which required little new field labor. That's the pattern I see most often: the headline "target better" lever covers its cost, and the quieter "capture more from what you already have" levers are what push the whole thing into solidly worth-it. A shop that bought the data only for cold targeting and ignored its own book would have called it a wash. Same tool, different conclusion, entirely because of which levers they pulled and whether they measured.

What to ask a vendor before you sign

The sales call is where you find out whether the ROI is plausible or whether you're about to fund someone's pipeline. Bring these questions, and weight the answers heavily.

• "Can I run a sample on streets I already know?" Non-negotiable. Pull the data for a block where you know the actual roof ages and storm history cold, and check whether the list agrees with reality. If they won't let you sanity-check a sample, that's a tell about data quality.
• "Where does your roof-age estimate come from, and how often is it refreshed?" Permit records plus aerial change-detection is stronger than year-built; year-built dressed up as roof age is weak. Stale imagery means stale estimates. Get specifics.
• "Is your storm data per-roof or per-ZIP?" A model that estimates hail and wind impact on the individual property is more useful than a flag that says a storm passed through the area. Both have a place, but they're priced and valued differently.
• "What's the contract term and the out clause?" Annual lock-ins change your break-even because you're committing the spend whether or not the first quarter works. Prefer a real trial window.
• "How does the list get into how I already work?" If it lives in a portal nobody opens, adoption dies and the ROI dies with it. Two-way sync into the CRM your reps already use is worth a premium because adoption is most of the battle.
• "What does the data NOT do?" A vendor who can articulate their own limits — roof age is a range, storm modeling is odds not proof, scoring is heuristics not certainty — is more trustworthy than one who claims their data is truth. Anyone selling certainty in this category is selling you something.
• "Show me how I'd measure ROI with your tool." If the answer is a vague case study about somebody else's shop, push back. You want to know how to isolate your incremental lift and your cost per win. A tool with a built-in results funnel that tracks actual-versus-estimate makes this concrete; a portal that just dumps a list leaves the measurement entirely on you.

How your baseline changes the whole calculation

I keep returning to baseline because it's the variable that flips the verdict more than any vendor feature. The same data, at the same price, is a great buy for one shop and a waste for another purely because of where they're starting. Here's the rough map:

The practical takeaway: don't ask "is this data good?" in the abstract. Ask "is this data meaningfully better than what I'm doing right now, for the levers I can actually pull?" A mediocre baseline makes ordinary data look brilliant; an excellent baseline makes excellent data look pointless. Your honest assessment of your own starting point is half the ROI answer, and it's the half no vendor can give you.

Where RoofPredict fits — and where it doesn't

Since I said I'd use one platform as a concrete example: RoofPredict is one option in this category, and I'll describe it the way I'd describe it to a friend, limits included. Its model scores roofs house-by-house by roof-age band plus the storms each roof has actually taken (it models hail and wind impact per property rather than just flagging that a storm passed through the ZIP), and it ships the things that drive the levers above — a ranked list of which roofs are due, tracked direct mail with proofs and delivery tracking, per-home homeowner reports/microsites/PDF+QR for warmer conversations, a canvassing app, a CRM/leads pipeline with two-way sync to a wide set of CRMs (HubSpot, ServiceTitan, JobNimbus, AccuLynx, Jobber, and others), a results funnel that tracks actual-versus-estimate and cost-per-win, and a claims/documentation module that does OCR and scope QA on the contractor's own documents within the public-adjuster boundary discussed above.

The honest limits, which apply to RoofPredict and to every competitor: roof age is a range, not an exact date; the scoring is roof-age and storm-exposure heuristics, not magic that knows the roof is failing; and a storm model is odds, not proof of damage. It is not a lead service and not a measurement tool. Anyone telling you their data is certainty is selling you something. The reason I find the per-roof storm modeling and the built-in cost-per-win tracking worth a look is precisely that they map onto the ROI levers and the measurement discipline this whole breakdown is about — the results funnel makes the "measure your payback honestly" step less of a manual spreadsheet chore. But evaluate it like you'd evaluate any of them: get a sample on streets you know, run a 90-day trial with a written threshold, and isolate the incremental wins. The framework decides; the brand doesn't.

The bottom line

Is roof targeting data worth it? It's worth it when three things are true at once: your sales motion can already convert (the data sharpens a working funnel, it doesn't build one), you have the labor to actually work a list, and you're disciplined enough to measure incremental lift instead of total volume. When those hold, the math frequently clears — especially once you pull the quieter levers (CRM reactivation and tighter supplement documentation on jobs you already have) and not only the loud one (cold targeting).

It's NOT worth it when you bought it expecting leads, when you're too small to feed it, when your baseline targeting is already excellent, or when you won't run the numbers. In those cases the subscription is the cheapest part of the money you'll lose.

Run the break-even on your own figures before you sign anything. Three incremental jobs a quarter is a low bar for most shops; the question is whether your lift, on your baseline, with your close rate, clears it — and the only way to know is to measure it like you mean it.

FAQ

How quickly should roof targeting data pay for itself?

Don't judge it before 90 days, because the first 30–60 days are onboarding ramp where you're paying full price at partial adoption and the numbers are noise. Set a written success threshold up front (for example, a target cost-per-win or a number of incremental jobs) and make a clean keep/kill decision at the end of a 90-day trial. For many shops the break-even is only a few incremental jobs per quarter, so a working program should clear it inside one to two quarters — if it hasn't after a full, well-measured quarter, the lift probably isn't there for your situation.

What's the difference between roof targeting data and a roofing lead service?

Targeting data tells you which doors to knock or which addresses to mail — it's fuel for outbound. A lead service hands you a homeowner who already raised their hand (often resold to several competitors). They're different categories with different costs and different ROI math. The most common reason people feel burned by targeting data is buying it while expecting leads; no amount of data quality fixes that mismatch.

Isn't roof age from a county assessor or Zillow good enough to target for free?

Year-built tells you when the house was built, not when it was last re-roofed, so a home built in 1995 that got a new roof in 2017 looks 'old' in those free sources and is actually new. Re-roofs are invisible to year-built data. Better targeting data tries to estimate actual roof age from permit records and aerial change-detection and gives you a range, not the build year. Free sources are a fine first filter but they'll send you to a lot of new roofs.

How do I calculate the break-even for buying roof data?

Add your total program cost for the period (subscription + fully-loaded labor to work the list + onboarding + any mail costs), then divide by your gross profit per won job — not revenue. That's your break-even job count. The critical step is counting only INCREMENTAL wins (the jobs you would NOT have gotten anyway), which you isolate with a hold-out crew or by tracking a clean efficiency ratio like jobs-per-100-knocks before and after.

What costs do people forget when budgeting for targeting data?

The subscription is usually the smallest piece. The big forgotten costs are the field/mail/dial labor to actually act on the list, the 30–90 day onboarding ramp where adoption is partial, data refresh to fight staleness, and the opportunity cost of rep hours spent on the list instead of referrals or repeat customers. Budget all five or your ROI estimate will be wrong in the optimistic direction.

When is roof targeting data NOT worth buying?

When your sales motion can't convert (data sharpens a funnel, it doesn't build one), when you're too small to field the labor to work a list, when your hand-picked baseline targeting is already excellent so the lift is small, when you bought it expecting inbound leads, or when you won't commit to measuring incremental lift. In those situations the subscription is the cheapest part of the money you'll lose.

Can targeting data help with insurance supplements on storm jobs?

Indirectly, and within a strict boundary. A contractor may use tools to document its OWN inspection, estimate, photos, and measurements, run scope QA, compare its own estimate to the carrier's internally, and track its submitted paperwork — and that can surface documented scope it didn't capture, which is real recovered revenue on jobs you already have. A contractor may NOT represent, interpret, or negotiate the homeowner's claim, advise on coverage or settlement, or use deductible-waiver messaging — that's licensed public-adjusting in most states. The homeowner files, the insurer decides, the contractor documents its own scope.

Does better targeting actually improve direct mail ROI?

It can, because mail ROI is brutally sensitive to response rate, and removing the not-due homes from a drop is supposed to raise that rate. Run mail break-even separately: cost per piece divided by (gross profit per job x close rate on responders) gives the response rate you need to break even. A smaller, better-targeted drop with a higher response rate often beats a big cheap blanket drop — but only if the targeting lift is real, which you confirm with tracked mail and per-piece delivery data, not by feel.

How do I measure ROI honestly instead of fooling myself?

Capture your baseline (jobs per rep, cost per acquisition, inspection-to-sale rate) BEFORE you start. Use a hold-out crew or a consistent efficiency ratio to isolate incremental wins. Keep the same definitions of cost and acquisition before and after so you're not quietly changing the denominator. Track the full funnel — contact rate, conversation-to-inspection, inspection-to-sale, cost-per-win — so you can tell where the value came from instead of crediting the data for wins you'd have gotten anyway.

What single metric best tells me if the data is working?

Cost per acquisition (cost per won job) is the cleanest single number because it folds cost and outcome together — but pair it with inspection-to-sale rate. If cost per win improved AND a higher share of the roofs you climbed turned into sales, the targeting is genuinely doing its job. If cost per win improved but inspection-to-sale didn't, your gain probably came from somewhere else and the data may not deserve the credit.