Do I need an attribution tool? The honest case for and against, from the vendors' own pages

What an attribution tool actually adds

An attribution tool is software that records visits and orders with its own tracking, joins them to the spend each ad platform reports, and applies one credit rule across every channel. Each ad platform can credit only the touches it can see: Google's data-driven attribution page says its model looks at interactions with Google's own Search, Shopping, YouTube, Display and Demand Gen ads, and Meta's attribution settings page counts conversions that follow an impression, a link click or an engagement with a Meta ad. A sale that followed both a Meta click and a Google click is one order in the store and a whole sale in each platform. The tool's job is to hold one record of that order and give the credit once.

The vendors describe the same three parts. Northbeam's pricing page lists independent, first-party multi-touch attribution across ad channels, view-through credit through a product it calls C+DV, and Apex, which feeds its attribution back into the ad platforms' delivery. Triple Whale's pricing page lists a Triple Pixel, first and last click attribution on its free tier, every model and lookback window on the paid tiers, a post-purchase survey, and marketing mix modelling and incrementality testing as add-ons on its top two tiers. Hyros's pricing page says it runs off existing UTMs, changes nothing about the ads, and shows its reporting inside Facebook's ads manager. Meta's own settings page describes the receiving end of Apex's loop as custom attribution: an advertiser shares attribution data from an external tool so that Meta's delivery optimises toward it.

What a tool does not change is that its output is still a model. Meta's Conversion Lift page lists "does my current attribution model align with the results of a controlled experiment" as a question only a lift test answers. Shopify's marketing reports help notes that its any click model allocates more credit than the orders received, which is what adding up the platforms' own claims does. A tool that credits more sales to your ads than Meta does has changed a report, not your revenue. Our guide to multi-channel attribution models works five orders through four models to show how far the credit moves.

When platform reporting is enough

Platform reporting is enough when every decision you will actually make sits inside one platform. Which Meta ad set to scale, which creative to retire, which Google keyword to pause: each platform answers those with its own attribution, and the figures are comparable within the platform as long as the attribution setting is the same. Meta's settings page warns that results cannot be compared across ad sets with different attribution models, and offers 1-day or 7-day click, 1-day view and 1-day engage-through windows for website and in-store conversions. Google says data-driven attribution is the default for most conversion actions, runs at any volume, and works best with at least 200 conversions and 2,000 ad interactions in a 30-day period.

The question platform reporting cannot answer is how much credit the other platform deserves. For a brand on one paid channel that question does not arise. For a brand on two, the cheapest check is not a tool but a lift test: Meta's Conversion Lift splits an audience into a test group that can see the ads and a control group that cannot, costs nothing extra, and, as a guide, needs a campaign started in the past year with USD 5,000 or more of spend and at least 500 conversions, prorated for campaigns longer than 90 days, plus a signal-quality requirement on the conversion data it is sent. Google's Conversion Lift runs the same experiment by users or by geography but is not available to every account; Google says to contact an account representative.

Two habits make platform figures safer to use. Google's attribution models page says a change of model counts conversions differently only from that day on, so pick a model and keep it for the comparison period. And track total revenue over total ad spend every week: it needs no model, and it is the one figure the platforms cannot inflate by claiming the same order twice. Our Meta vs Shopify ROAS guide walks through that reconciliation on real report exports.

What a DIY UTM setup can and cannot tell you

A UTM setup tells you which tagged link the shopper's browser last followed into your store before an order, on that device, as far as the store's cookie could see. Google Analytics' URL builder guide asks for utm_source, utm_medium and utm_campaign on every link, warns that values are case sensitive (Meta and meta are two rows), and says a missing parameter shows as (not set). Shopify's marketing reports put UTM-tagged sales in a Sales attributed to marketing report and offer five models over the referrers they recorded: last non-direct click (the default), last click, first click, any click and linear. Both are free, and both are enough to rank channels by the last click they can see.

The earlier version of this post called the result a perfect, 100% accurate last-click model. Shopify's own page says otherwise. Sessions and visitors are counted with cookies stored on the shopper's device, so a click on a phone and an order from a laptop are two visitors with two paths. A session ends after 30 minutes of inactivity. If a visitor does not buy within 30 days of a session, the first-interaction referrer resets to whatever comes next. The referrer name is listed as N/A when the browser has Do Not Track on, a proxy or firewall strips the referrer, or the link was a shortened URL. And a click is all it records: an ad that was seen and acted on later through a search shows up as search. Google Analytics adds one more rule, from its attribution guide: direct visits get credit only when the whole path was direct.

A UTM setup is accurate about the thing it measures. The mistake is reading it as a count of what the ads caused. For that, the table below sets each option against what its vendor documents.

The minimums come from Google's data-driven attribution page, Meta's Conversion Lift page and Meridian's data guide, which calls two years of weekly data a general rule of thumb for a model by region and three years for a national one. The prices are from each vendor's pricing page on the date in the caption and are read in the next section.

What attribution tools cost today

On October 22, 2026, one of the three attribution vendors this post named in July publishes a price. The earlier version said the tools typically start at $500 a month and climb to $1,500 or more; none of the three pages says that, and two publish no figure at all. The table records what each page shows.

Two things on those pages matter more than the figures. Northbeam places Starter under $1.5 million a year of ad spend, which is $125,000 a month, and Growth under $200,000 a month with an agency attached: the vendor itself is not pitching a brand at $10,000 a month. And a free tier with first and last click attribution is close to what Shopify and Google Analytics already give you; the paid step is multi-touch, view-through and the feedback loop into ad delivery.

Spend levels where a tool starts to pay

A tool starts to pay when the extra contribution it produces each month at least equals its fee. The method below is this article's, and every figure in it is invented to show the arithmetic; a brand should put in its own fee, margin, spend and return. The assumptions are stated so they can be changed:

1. Ad spend stays the same. The tool pays by moving budget to better ads, not by spending more.
2. Extra revenue counts at contribution margin. The example uses 50%, meaning half of each extra sale is left after product, shipping and payment costs. Use your own figure.
3. The fee is the lowest published figure. $1,500 a month for Northbeam Starter, and $3,500 for Professional as the second column. Northbeam shows a period only beside the Professional figure, so the example reads Starter as monthly too; confirm that on the call. Staff time to read the tool is left out, so the real bar is higher.
4. Return before the tool is 3.0x. Revenue from ads is three times spend, so $10,000 of spend returns $30,000.

With those inputs, a $1,500 fee needs $1,500 divided by 0.5, or $3,000, of extra monthly revenue from the same spend. A $3,500 fee needs $7,000. The table divides each by the revenue the spend already returns.

The lift has to be caused by a decision the tool enabled and you would not otherwise have made. A tool that credits more sales to your ads than Meta does, which is the comparison the vendors' pages lead with, has not lifted revenue; it has moved credit. The only gain that pays the fee is a reallocation that the platform's own reporting would not have shown you. At $10,000 a month, that reallocation has to find a tenth more revenue every month, from a budget where most decisions sit inside one platform anyway. At $100,000 a month across two or three channels, one percent is a plausible return on knowing which channel to feed.

The same method applies to the free options. A lift test costs no fee, so its bar is the time to set it up and the sales the control group forgoes. Google Analytics and Shopify cost nothing extra, so every click-based question they answer is a question a tool has to answer better, not merely answer.

Questions to ask before buying

The questions below come from what the vendors' own pages leave open, and an answer to each should be in writing before a contract is signed.

1. Which model is the default, and can I see first click and last click beside it? Triple Whale's free tier offers first and last click only; Northbeam and Triple Whale's paid tiers list every model. A tool that shows one model hides how much the credit depends on it.
2. What happens to an order the pixel never saw? Shopify's cookie rules apply to any pixel: a buyer who switched devices, blocked the referrer or came back after a quiet month is a gap. Ask how the tool matches those orders and what share of orders it matched in a trial.
3. Does it feed data back into my ad accounts, and what changes if it does? Northbeam's Apex feeds its attribution back into the ad platforms' delivery, and Meta's custom attribution option optimises delivery toward whatever is sent. That is a change to how your ads are delivered, not only to a report.
4. What is the price, in writing, for my spend? Northbeam's FAQ says its plans start at the rates shown. Hyros sizes its price to the revenue it tracks. Triple Whale publishes none. Ask what the fee does when spend or revenue grows.
5. How long is the contract? Northbeam's pricing FAQ has an entry on contract length and Triple Whale's one on monthly versus annual subscription. Read the answers before the demo, not after.
6. Can I export the touchpoint data? Northbeam lists granular touchpoint export on its Enterprise plan, and Triple Whale lists data warehouse sync as an add-on on every paid tier. Without an export, the model and the history leave when the contract does.
7. Who will read it every week? A tool that changes no decision is a report. Name the person and the decision before the trial starts.

Where overads fits

overads is an AI marketing agent for small teams: it drafts posts, reports and recommendations from your connected accounts and waits for a person to approve them. It is not an attribution tool. On the ad side, overads reads Meta, Google, X and Snapchat ad accounts read-only and stores each platform's figures as that platform attributed them. For Meta, Google and Snapchat we checked this in its code on October 22, 2026, before X Ads went live: the Meta sync names no attribution window and so receives Meta's default, Google's conversions arrive under each conversion action's own model and are rounded to whole conversions, and Snapchat is account-level spend only. The analytics API returns spend and results per day, split by account, platform and currency, with a missing figure as null rather than 0, so a quiet day and a day with no reading are told apart.

What overads does not do:

• It has no attribution model of its own, so it cannot split one sale between Meta and Google Ads or credit a view.
• It does not read store orders, so the blended ratio stays a spreadsheet job: the media side from the API, the sales side from the store.
• It does not run lift tests or marketing mix models.
• It does not change ads. Access is read-only today, for the reasons in our guide to why many ad tools can read campaigns but not change them, and any change it suggests is applied by a person in the ad platform.
• It does not read LinkedIn, Pinterest or TikTok ad accounts.

A brand on two or more paid channels that has cleared the break-even bar above is better served by one of the tools in the pricing table, or by a lift test and a marketing mix model fed from a warehouse. A brand below it gets the within-platform view, with each platform's figures kept as reported, from overads' Meta ads reporting for DTC operators and the rest of its DTC operator workflows. For how reporting tools compare on Meta, Google and Snapchat, see cross-platform ad reporting tools compared, and for the metric definitions that differ by platform before any attribution is applied, our guide to normalizing cross-platform ad metrics.