How to read Instagram ad rates before you set a budget

Selzee's mid-October 2026 page states the constraint in one place: "Instagram has no separate price list. It sells its ads through the same auction as Facebook: you set the budget, and the auction sets the price of each impression."

CPM, the cost of 1,000 impressions, and CPC, the cost of a click, are the prices that auction produced. Adwize, writing in late October 2026, reports that Meta usually charges per impression, so a CPC is spend divided by clicks. The same page reports that the minimum depends on business vertical, budget type, buying type, bid strategy, optimisation, currency and schedule. Adwize is relaying Meta's minimum-budget page. Those factors describe that minimum, not the makeup of every published CPM.

The USD figures do not describe the same ads

Every figure below is in US dollars, as published. A dollar figure is not a sterling price, and these rows do not establish a UK rate or an India rate.

| Figure | Where it is published | Population and period | Limit | | --- | --- | --- | --- | | $2 to $6 CPM for better-performing campaigns; $0.40 to $0.70 per click counting all clicks; $0.50 to $1.20 per click to a website | Shaazford, quoting WordStream's 2026 ranges | WordStream 2026 ranges, as quoted by Shaazford | The $2 to $6 CPM applies only to better-performing campaigns. The two click ranges count different clicks. | | $6 to $11 CPM; $0.40 to $1.70 per click; $0.01 to $0.06 per engagement. AdNabu also reports around $38 per acquisition | AdNabu, about 20 July 2026 | The CPM, click and engagement ranges are cited to Cropink. The acquisition figure is AdNabu's, not part of that Cropink citation. AdNabu also says most small businesses spend $300 to $500 a month to start | Cropink's period and method are not identified with these ranges. The $0.40 to $1.70 range does not identify the click type. | | $15.26 CPM; $1.31 CPC; CPE $0.067; CPL $11.31 | Hootsuite's own guide, updated about mid-October 2026 | Prose says 2025 averages. The table heading says 2026 Instagram cost benchmarks | The year label disagrees with itself. | | About $9.46 Instagram CPM, beside a Facebook CPM of about $8 to $9 | Travi, 29 September 2026, attributed to Hootsuite | Described as a broad 2026 benchmark | Conflicts with the $15.26 on Hootsuite's own page. This article does not choose between them. | | $15.06 CPM; $38.99 per purchase. Prior year: $13.30 and $37.80 | Selzee, page about mid-October 2026 | Column labelled Meta median, August 2025 to July 2026. Not labelled Instagram-only | Gidds Media attributes the same two numbers to Triple Whale's 2026 report for online stores or online brands. |

Shaazford quotes WordStream's 2026 ranges as $0.40 to $0.70 per click when all clicks count, and $0.50 to $1.20 per click to a website. A click that only likes or expands the ad is a different step from a click through to the site. The material used here does not establish the click type behind AdNabu's or Hootsuite's click figures.

AdNabu reports around $38 per acquisition, and Selzee reports $38.99 per purchase. The figures sit near each other, but they come from different pages, and this article has not established a shared method.

Gidds Media passes on the same pair of numbers, describing them as a cost to show an ad to 1,000 people, while Selzee's column is a cost per 1,000 impressions. Read Gidds as the same reported chain, not a second measurement.

Averaging these ranges invents a rate

A blend of $2 to $6 and $15.06 describes no row in the table. The $2 to $6 band is limited to better-performing campaigns in the WordStream ranges Shaazford quotes. Gidds Media ties $15.06 to online stores or online brands, on Meta rather than Instagram alone.

Travi's $9.46 and Hootsuite's $15.26 should not be pasted into a Hootsuite range either. One number is on Hootsuite's page. The other is another site's citation of Hootsuite. If you cite Hootsuite, cite the page, and keep the 2025 and 2026 label problem beside the number.

Novoads reports that in its second quarter 2026 results, Meta reported that its average price per ad rose 12% year on year. That is Novoads retelling Meta's Q2 2026 results. Price per ad is a global average across Meta's surfaces, so it is a different measure from an Instagram CPM. The move from $13.30 to $15.06 in Selzee's table points the same way and measures something else. Do not add the 12% to the median, or treat the median's rise as that 12%.

CPM and CPC are what the auction produced

Adwize moves the buying question to what the budget buys: impressions, clicks and sales, and the cost of each.

Travi lists what moves the cost of a reach, a click, a lead or a sale: competition, audience, objective, creative quality, seasonality, and the signals available to delivery. Seasonality is named, not quantified. The August 2025 to July 2026 median covers twelve months, so do not use it as a November price.

The reported score that sorts the auction

A late September 2026 page on mynameiskalam.com relays an auction description and dates the wording to Meta's 2019 post, "Good Questions, Real Answers: How Does Facebook Use Machine Learning to Deliver Ads?", which that page labels historical. On that account, the ads shown are those with the highest total value, a combination of advertiser value and ad quality, and advertiser value equals the bid multiplied by the estimated action rate. The same page says quality models use feedback from people who view or hide the ad, plus assessments of low-quality attributes. Use it as a historical, second-hand account of how bid, estimated action rate and quality sort an impression.

Klipio writes the score as an approximate sum: bid times estimated action rate, plus ad quality. It also says estimated action rate and ad quality together are what Meta calls ad relevance. Klipio calls that sum an industry teaching model and says Meta has not published the exact formula. Clarigital prints a different version, multiplying by ad quality instead of adding it. Because those secondary pages disagree, neither formula is treated here as Meta's specification.

On that reported account, the bid is only one part of the score. A weaker estimated action rate, or weaker quality signals, can lose the impression or make the same impression cost more.

An Instagram-only CPM is a weak input for mixed delivery

If delivery can use Facebook and Instagram together, the Instagram line in a benchmark table is not the price of the campaign.

Selzee reports a hypothetical Meta Advantage+ placements example: Instagram at $5 per result, Facebook at $3. Switching Instagram off gets 8 results for $26 instead of 9 for $27. In that example the dearer placement still added a result and cutting it saved $1, but cost per result rose from $3 to $3.25 because the remaining spend moved to Audience Network. Shaazford reports Meta's placements guidance that adding placements does not increase the cost of the ad, and says those rules were checked on 7 October 2026.

That reported example is one case. Use it as a reason to avoid budgeting a mixed ad set from an Instagram-only CPM.

A worked cost per order from the Meta median

The rates in this section are assumptions. They are not Triple Whale rates, not Selzee rates, and not Meta rates.

Start from the Meta median CPM of $15.06 for August 2025 to July 2026. Assume a 1% link-click rate: 10 link clicks per 1,000 impressions. Cost per link click is $15.06 divided by 10, which is $1.506.

Assume 2% of those link clicks become a purchase: one purchase per 50 link clicks. Cost per order is 50 times $1.506, which is $75.30.

Order cost moves inversely with the link-click rate and the purchase rate, not in a straight line. At a 0.5% link-click rate the click cost doubles, and so does the order cost, if the purchase rate stays at 2%. At a 4% purchase rate the order cost halves, if the click rate stays at 1%.

Set that cost beside the margin

The store in this section is invented. Average order value $90. Contribution margin 40% after cost of goods, shipping, payment fees and returns, and before ad spend. Break-even cost per order is $90 times 0.40, which is $36.

$75.30 is above the $36 break-even. The published Meta median purchase cost, $38.99 for the same August 2025 to July 2026 window, is also above $36. Under these invented economics, a CPM from Shaazford's $2 to $6 band for better-performing campaigns still needs a click rate and a purchase rate that bring the order cost under $36.

A different order value or a different margin produces a different ceiling. How that ceiling sits next to MER is a separate read, in Conversion metrics that actually matter for paid acquisition decisions.

A test budget from the reported rules

These budget rules come from vendor pages. They were not checked against Meta's primary text for this article. Treat them as a sizing sketch, then confirm them in the ad account's own help text before you rely on them.

Selzee reports that if you use a cost-per-result goal, the daily budget should be at least five times that goal. On that report, the floor also depends on vertical, budget type, bid strategy, optimisation event, currency and schedule. Applied to the invented $36 ceiling, five times $36 is a $180 daily budget. That sizes a test of a $36 goal. It does not make $36 the right goal for your catalogue.

Novoads, in early October 2026, and Salesio, around August 2026, report a different starting line: at least US$5, for more than six days, so delivery can find suitable people. Novoads also reports at least seven days for the system to learn. Salesio is explicit that the US$5 line is a starting-budget recommendation, not a guaranteed price for an outcome. US$5 does not meet a five-times floor on a $36 goal. Do not rewrite it as £5. The reported minimum depends on currency, and Salesio states this figure in US dollars.

Both Novoads and Salesio report the pacing cap. On Salesio's wording, daily spend can reach 75% above the promotional daily amount, and weekly spend will not exceed seven times the daily budget. Applied to the $180 daily budget in the example above, one day can reach $315 and the week can reach $1,260. If cash is tight, do not plan that week as seven even days of $180.

AdNabu's Cropink citation, $300 to $500 a month for most small businesses starting out, answers a smaller question, and Cropink's method was not verified. That month does not fund a $1,260 week.

Before you raise the daily figure past the test, use the checks in What to check before increasing Meta ad spend.