Programmatic build economics: will a large boiler page set pay for itself?

The three numbers that decide everything

You need three numbers before any other calculation means anything.

1. Gross profit per order or job. Revenue minus the direct cost of parts, equipment and labour. For a parts seller, maybe $40 to $200 an order; for an installer, often thousands per replacement.
2. Conversion from visit to order or call. Pull it from your analytics and call records. If you don't have it, that's the first thing to fix.
3. Total cost of the page set. Build cost plus monthly costs over the period you're judging.

Without the first number, you can't tell a good page set from an expensive one. Our pricing page gives the cost side; your books give the profit side.

Cost per useful page

The most honest single measure of a programmatic build is cost per useful page: total cost divided by the number of pages that actually earn clicks.

Here is a worked example using published HVAC.Agency prices:

• Enterprise website: $12,500 one-time, plus $449/month
• Growth plan: $4,498/month
• First-year total before GST/HST: $12,500 + ($4,947 × 12) = $71,864

Say the site publishes 1,500 pages in its first year. After 90 days per batch, you find:

These shares are illustrative, not a forecast; your own reporting replaces them. On these numbers, cost per published page is about $48, but cost per useful page, one with clicks, is about $240. That second figure is the one to compare against your gross profit.

Note that the plan also covers ads, reporting and other work, so this method overstates the cost of the pages alone. It's still the right conservative test.

Break-even for a parts seller

For a parts seller, break-even is about order volume.

Continue the example. First-year cost with 5% GST, as in BC or Alberta, is $75,457.20. If your average gross profit per order is $90:

• Orders needed to break even: $75,457.20 ÷ $90 ≈ 839 orders a year, about 70 a month.
• If 1.5% of visits become orders, you need about 55,900 visits a year, or around 4,700 a month.
• Spread over 300 useful pages, that is about 16 visits per useful page per month.

Now you can ask a concrete question: is 16 visits a month realistic for a typical part page in your catalogue? Your Search Console data on existing part pages answers it far better than any agency promise.

BoilerPart.ca, our reference build, runs 2,000+ programmatic pages and grew organic traffic 180%. That shows parts traffic can grow at scale. It does not tell you your conversion rate or margin, which is why the sum above uses yours.

Break-even for an installer

For an installer, the numbers look different: fewer pages, far fewer conversions needed, and much higher profit per job.

Take an installer with 10 city hubs, a Professional website at $6,500 plus $249/month, and a Starter plan at $2,998/month.

• First-year cost: $6,500 + ($3,247 × 12) = $45,464 before tax, or $47,737.20 with 5% GST.
• If a boiler replacement brings $3,500 gross profit and a service call $150, then 14 replacements alone cover the year.
• If one in five calls from city pages becomes a replacement, that is about 70 qualified calls a year, or fewer than one call per hub per month.

This is why city hubs often pay back faster than large catalogues. The economics favour a few strong pages with high-value conversions over many pages with small ones.

When to stop adding pages

Each new batch should earn its place against the previous one. Stop or slow down when:

• the indexed share of new batches falls well below earlier ones
• new pages earn impressions but almost no clicks, which often means they compete with your own stronger pages
• your remaining rows only pass the data threshold because the line was lowered
• maintenance, such as supersessions and discontinued models, starts eating the time meant for new pages

A practical rule: compare the share of pages with clicks in your latest batch against your first. If it has halved, the next batch probably isn't worth building until the data improves.

The quarterly review

Three checks, every quarter, keep a page set honest.

1. Cost per useful page, by page type. Is it rising or falling?
2. Gross profit traced to the page set, from orders, calls and texts. Does it exceed cost, or is it on track to?
3. Pages to prune. Pages with no impressions in 12 months should be improved, merged or removed; see our scaled-content risk audit.

Write the three numbers down each quarter. After a year you'll have the clearest picture of your marketing you've ever had.

Common mistakes in the sums

• Using revenue instead of gross profit. It flatters every result.
• Counting all pages as useful. Only pages with clicks do the work.
• Ignoring GST/HST and monthly fees. One-time prices are only part of the cost.
• Judging too early. Large sites are revisited gradually; a month tells you little.
• No call tracking. For installers, most conversions are calls and texts, and without tracking they vanish from the sum.

If you'd like to run these sums on your own numbers with us, the 7-day free trial includes an audit and an estimate of how many pages your data supports. Our services overview explains what each piece of work costs in time. Call, text or email with your margin and your current traffic, and we'll work it through.