The IPQ Field Guide

Architecture marketing benchmark report 2026

Someone has quoted you a cost per lead. Follow that number back to where it came from and the trail usually ends at dentists.

Arseniy VolosevichFounder, IPQmedia·Updated Sep 11, 2026·18 min read

The largest public Facebook advertising benchmark set covers fifteen industries. Architecture is not one of them. Neither is construction. So when a firm gets told what a lead "should" cost, the number is almost always borrowed from a category that sells a $9,000 job to someone who decided this week, then quietly applied to a practice that sells a $2 million build to someone who has been thinking about it for three years.

This report is our attempt to fix that for our own corner of the market. Every figure below comes out of ad accounts we run for US architecture and design firms, re-pulled from the reporting tables on 11 September 2026. Aggregate and anonymized, with the sample size attached to each claim, including the claims that make us look expensive.

Second edition. The first ran in July on sixteen accounts and ten months of spend. The window is now a full twelve months across twenty-one accounts, which was enough to overturn two of the July findings. Both corrections are below in their own section rather than quietly swapped out.

Why nobody has published this before

We went looking for prior art first, on the assumption that somebody must already benchmark paid social for architecture firms. Nobody does. LocaliQ and WordStream publish the most-cited Facebook benchmark study there is, and we re-read its industry list on 11 September 2026. Fifteen rows, from dentists to restaurants. No architecture row, no construction row, no design row.

The trade research that does exist for the profession, from SMPS and Deltek and the AIA, covers what firms spend as a share of revenue and how they win work. It does not publish what a lead costs.

The nearest published anchor is Home and Home Improvement at $41.26 per lead, against an all-industry average of $27.66 for lead campaigns (LocaliQ and WordStream, Facebook Advertising Benchmarks, last updated 24 October 2025). Those are the numbers a principal is unknowingly compared against. Neither was measured on anyone selling architectural services.

The benchmark table

Quartiles, not averages. One loud account can drag an average anywhere, so each row below is built from the 21 accounts as 21 separate results and then cut at the 25th, 50th and 75th percentile. If your own number sits at or under the median column, you are in the better half of the accounts we run.

Two click rows, because the word "click" is doing something sneaky. Meta's headline cost per click counts every click on the ad, including reactions, comments and taps on the firm's profile photo. The click that can actually turn into a lead is the one that opens the ad's destination.

In these accounts the median firm paid $2.11 for the first kind and $3.60 for the second, so the number worth planning against runs about 70 percent higher than the number the dashboard shows you first. Our July edition quoted the all-clicks figure without saying so, which is the kind of thing a benchmark should be explicit about.

The full cost per lead range across the 21 accounts was $12.58 to $112.94. That is a factor of nine between the cheapest and the most expensive practice buying leads in the same country, in the same year, on the same platform.

That spread is the actual finding. There is no single architecture cost per lead to hit, and any agency quoting you one number with a straight face is quoting you a median at best.

What a normal month looks like

Across 42 full calendar months of account activity, the median month was $1,453 of ad spend producing 22 leads. The quieter quarter of months came in at $1,141 and 18 leads. The busier quarter ran $2,534 and 40 leads. That is the honest shape of a small firm's advertising month, and it is a long way from the volume language used in most pitches.

Twenty-two leads a month sounds thin until you price it against the work. One signed residential project at a typical 8 to 15 percent fee usually pays for a year of that spend several times over, which is the whole argument for running a lead generation system at all rather than waiting for the phone.

Leads arrive faster than firms expect, then stop for days

Two numbers that sit uncomfortably next to each other. The median account got its first lead one day after launch, 17 of the 21 got one inside the first week, and the slowest took 12 days. Every account in the sample eventually produced leads. So the machine tends to prove itself quickly.

Then, across 1,839 account days of live spend, 41 percent produced no leads at all. Nearly every account we run has weeks with a two day silence in the middle of a perfectly healthy month. Firms who check the dashboard daily read that silence as failure and start turning things off, which is the single most expensive habit in this whole business.

July put that figure at 47 percent on ten months of data. More accounts and more months moved it six points in the firms' favor, and it is still two days in five.

Two things we got wrong in the July edition

A benchmark that never contradicts itself is a benchmark nobody is re-running. Two of the July findings did not survive a sample a third larger, and the shape of each error is more useful than the original claim was.

Cost per lead does not reliably climb as an account ages. July reported a drift from about $41 in month one to about $71 in month four, and said the drift survived correcting for seasonality. On twelve months it mostly does not.

Indexing every account month against what the whole sample paid in that same calendar month, months one to three sit at 1.00 times the sample (27 account months), months four to six at 1.03 (15 account months) and months seven and later at 1.29 (10 account months, across three accounts).

The early part of the July curve turned out to be an artifact of when accounts launched. Most of the newest ones started in a cheap stretch of summer, which flattered month one and made everything after it look like decay.

What is left is a late-tenure signal on three accounts, which is not enough to plan a budget around. Creative fatigue and a thinning local audience are still real and still worth getting ahead of. We are no longer willing to tell you they arrive on a schedule.

Daily budget does not predict cost per lead either. July found the accounts spending under $40 a day had the lowest cost per lead. In this cut they have the highest, at a $64 median against $37 for the $40 to $80 band and $55 above it. A relationship that flips direction in seven weeks was never a relationship. Spend level is not the lever, so the honest version is that we cannot find one in this data.

Spring is the expensive quarter

This is what the twelve month window bought, and we could not have published it in July. Compare each account month against that same account's own average for the year, so a firm's local market and offer cancel out and only the timing is left.

April through June runs at 1.12 times an account's own year (16 account months). October through December runs at 0.89 (10), January through March at 0.87 (11), July through September at 0.92 (15). The same firm, the same ads, the same county, pays roughly a quarter more for a lead in late spring than it does in the first quarter.

Our read, and it is a read rather than a proven mechanism: spring is when everyone in the home and building trades turns their advertising on, so an architecture firm is bidding against remodelers and builders and landscape designers for the same attention.

The practical version is unglamorous. Do not launch a new account in May and judge the channel on it, and do not cut budget in January because it looks quiet when it is actually the cheapest month you will get. The wider seasonal pattern in buyer demand itself is mapped in architecture client demand by metro.

Lead forms against landing pages

Splitting campaigns by where the lead was actually captured, 39 campaigns across 20 accounts ran on Facebook's own lead forms and produced 1,867 leads at $40.55 blended. Nine campaigns across 7 accounts sent traffic to a landing page and produced 439 leads at $70.18 blended. Across the whole sample, 78 percent of leads arrived through an in-platform form.

So in-platform forms typically come in around 40 percent cheaper per lead, and the gap is wider than the 07-26 cut showed. That is not an argument that landing pages are wrong, and the volume difference here is real enough to say loudly: nine campaigns is a small sample. A landing page also asks for more effort, which tends to hand you a better conversation with fewer people.

Firms with strong web pages built to convert rather than to impress often do better on the page route than this table suggests. The full method behind both routes is in Meta ads for architects.

One campaign does most of the work

Restricting to the nine accounts that ran three or more campaigns, so the number is not just an artifact of accounts that only ever ran one, the single best campaign produced 84 percent of that account's leads at the median. The weakest concentration in that group was still 46 percent from one campaign, and the strongest was 97 percent.

Across the whole sample, 33 percent of spend went to campaigns that were not the account's best one, and 8 of 59 campaigns produced no leads at all on $2,449 between them. Most of that spend bought the search rather than the result. You cannot know which campaign is the one until several have run.

That ratio is the practical case for volume in production and simplicity in execution. You are not looking for a set of good campaigns, you are looking for the one that works, and the way to find it is to put up several and let the losers die quickly. It is also why a steady supply of raw footage and project photography matters more than any single polished asset.

Method note for anyone comparing editions: July reported this at the individual ad level and found 56 percent of ads produced zero leads. Our current export holds performance at campaign level only, so this edition reports campaigns, and we have not restated the July figure as though the two were interchangeable.

From lead to booked consultation

Of 2,419 leads, 742 became a booked consultation inside the client's own advertising window. That is 31 percent pooled, at about $159 of ad spend per booked consultation. Per account, the median was 32 percent, the better quarter cleared 45 percent and the weaker quarter came in under 15 percent.

Read that spread carefully, because part of it is measurement rather than performance. Some firms book consultations in systems we can see and some book them over the phone into a calendar we cannot, so the low end of that range mixes slow follow-up with incomplete tracking.

We are also still not publishing a cost per consultation that actually showed up, and we checked again for this edition rather than repeating the July line. Of 1,023 booked consultations in the tracked calendars, only 48 carry a recorded outcome, and 18 of the 24 calendars record none at all. A show rate built on that would be a number about our record-keeping. When it is solid it goes on this page.

What the spread does establish is where the money leaks. A firm at 15 percent and a firm at 45 percent are usually buying the same leads at the same price, and the difference is speed of response and who picks up the phone, which is the part of the follow-up system nobody wants to own.

Residential and commercial read differently

Ten of the 21 accounts carry a segment on their record, seven residential and three commercial, and this is the first edition where the commercial side has enough accounts to say anything at all. In July it was one.

The residential accounts paid $39.92 per lead and booked 34.8 percent of those leads, landing at $114.65 per booked consultation. The commercial accounts paid $53.43 per lead, 34 percent more, on a $63.02 cost per thousand impressions against $51.45. Then they booked 43.2 percent of those leads and landed at $123.63 per booked consultation. Fewer, more expensive, and considerably more serious once they arrive.

The two segments converge at the metric that matters. A commercial lead costs a third more and earns more than a third back, because a materially higher share of them turn into a meeting. A firm reading only the cost per lead column will conclude commercial work is expensive to market, and the booking column says otherwise.

If you design homes, expect the shape in this report: a cheap first week, a $28 to $71 cost per lead band, a lot of unqualified curiosity mixed in, and homeowners who need several conversations before they are a project. The detail work for that segment sits in marketing for residential architects.

If you design for developers, owners and businesses, expect a higher cost per lead, a longer evaluation, and a better show of intent per inquiry. All three commercial accounts here ran a specific offer aimed at someone comparing firms rather than a soft form, which fits what we see in marketing for commercial architects: the same engine aimed at a buyer who is evaluating rather than dreaming.

Caveat stated plainly, because three accounts is three accounts. Treat the commercial figures as a direction, and re-read this section when the next edition lands.

How these numbers compare to the published ones

Our median firm pays $2.11 a click on the same all-clicks basis the published studies use, where the published all-industry Facebook figure is $1.92. Our median account converts 9.1 percent of link clicks into leads where the published click-to-lead figure is 7.72 percent. So the ads themselves perform normally. Our median cost per lead of $50 sits well above the published all-industry $27.66 and above the $41.26 for Home and Home Improvement.

Some of that gap is time. Facebook lead costs in that same published dataset rose 21 percent year over year, from $22.87 to $27.66, and their window closed well before ours runs to September 2026.

Some of it is the market. An architecture firm advertises to one or two counties, not a nation, and the pool of people about to commission a custom building in your county this quarter is small. And some of it is simply what a serious lead costs when the thing being sold takes a year to sign.

For scale on the other side of the ledger, the same body of research puts the average Google search cost per lead at $70.11, and a click from someone typing "architect near me" into Google runs around $9 before anyone fills in anything. Our read on that trade sits in Google Ads for architects, and the wider channel picture is mapped in the architecture marketing guide.

A few honest answers

How many leads should an architecture firm expect from Facebook ads?

In our accounts the median full month produced 22 leads on $1,453 of spend, with the middle band running 18 to 40 leads. Expressed per dollar, the median account got 19.8 leads per $1,000 and the best quarter got over 35.

If someone promises you 100 leads a month at a small firm's budget, ask what they count as a lead and what share of them booked anything.

How long does it take for architecture ads to start producing leads?

Faster than most principals expect. The median account in this sample saw its first lead one day after launch and 17 of 21 saw one within the first week, with the slowest at 12 days.

The harder wait is the one after that. Judging a campaign inside its first two weeks of volume is how firms talk themselves out of a channel, and 41 percent of live account days produced no leads at all, so a quiet stretch means very little on its own.

Why did my cost per lead go up after a few months?

Check the calendar before you blame the account. Our July edition said cost per lead climbs with account age, and a full year of data mostly took that back: once every account month is compared against what the whole sample paid in the same calendar month, the first six months are flat.

What does move is the season. The same account typically pays about a quarter more per lead in April through June than in January through March. Creative fatigue and a thinning local audience still bite, and new creative and a wider radius are still the fix.

Are Facebook lead forms better than a landing page for architects?

Cheaper per lead, typically. Lead form campaigns in this sample came in at $40.55 per lead against $70.18 for landing page campaigns, and 78 percent of all leads arrived through a form.

Whether cheaper is better depends on what happens next, since a form that takes eight seconds to complete sends you people who invested eight seconds. Most of our accounts run forms as the volume engine and treat the landing page as the higher-intent route.

What does a booked consultation cost an architecture firm?

About $159 of ad spend at the pooled rate, and $148 for the median account, with the middle half of accounts between $105 and $240. Residential accounts landed at $114.65 and commercial accounts at $123.63.

That is the number worth holding. A lead is an intermediate metric and a consultation in the diary is the thing you can actually sell into. Price it against what one signed project pays you and the channel usually stops looking expensive.

How this report was built

Sample: 21 US architecture and design firm Meta ad accounts managed by IPQmedia, all spend between 15 September 2025 and 11 September 2026, re-pulled from our reporting tables on 11 September 2026. Totals: $117,656 of ad spend, 2,151,042 impressions, 81,038 link clicks, 2,419 leads, 742 booked consultations. Two construction contractor accounts in the same reporting tables were excluded, because they are a different profession and a different buyer.

Method notes. Per-firm figures treat each account as one observation and report quartiles, so a single high-volume account cannot move the median. Pooled figures divide total spend by total results and are labeled as such wherever they appear. Booked consultations are counted only inside each account's own advertising window.

Click-based rows exclude 25 account days, out of 2,091, where the export reported more link clicks than total clicks. That is a reporting artifact rather than a result, and leaving it in would have pulled the pooled cost per link click down by more than half. Those days remain in the spend and lead figures, which are unaffected.

The tenure analysis normalizes each account month against the whole sample's cost per lead for that calendar month, which is how we separated account age from seasonality. The seasonal analysis does the reverse, normalizing each account month against that account's own annual cost per lead, which is how we separated season from account mix. Both restrict to account months with at least 20 days of spend and at least 5 leads.

Limits, stated up front because a benchmark without them is marketing. Twenty-one accounts is a small sample and the commercial segment rests on three of them. Only three accounts have run long enough to say anything about month seven and beyond, which is exactly why the July tenure finding was retired rather than restated.

The rules we apply to every figure on this site, our own and everyone else's, are written down in our editorial standards, along with a dated log of the corrections we have had to make.

Show-rate and closed-project data are not clean enough across accounts to publish, so nothing here claims a cost per signed project or a cost per consultation held. All of it is Meta only, since that is what we run. Numbers move, and we refresh this page as the sample grows.

Third-party figures cited: LocaliQ and WordStream, Facebook Advertising Benchmarks (last updated 24 October 2025, read again 11 September 2026) for all-industry and per-industry cost per lead, cost per click, click-through rate and click-to-lead rate, and for the 21 percent year over year rise from $22.87 to $27.66; Search Engine Land, September 2025, for the $70.11 Google cost per lead. Agency pricing ranges and the firms buying these keywords are covered in our agency comparison.

Use it as a mirror, not a scoreboard

If you are running ads now, the useful move is to find your own row in the table above and look at the one metric where you sit in the costlier quarter. That is usually where the next month of work is.

If you are not running ads yet, the number to hold onto is the $148 per booked consultation, set against what one signed project pays you, because that ratio is why firms keep the channel on through the quiet weeks. To turn a budget you are considering into the leads and consultations these figures imply, the architecture marketing budget calculator runs the arithmetic on this same dataset. The alternative is waiting for referrals and calling the wait a strategy.

Running that engine for US architecture firms is what we do all day, and we charge for it when a qualified consultation actually shows up. If you want your own row to sit in the better quarter of this table, that is the conversation.