Keitaro vs Measured

What each one actually is

Keitaro is a tracker you install on your own server. You route paid clicks through it, and it stamps every click, records the conversion on a box you control, matches it back to the source through server-to-server postbacks, and reports across GEO, device, browser, offer and lander. Its own homepage frames the job plainly: see exactly how your marketing efforts drive sales, from one source of truth. There is no SaaS vendor holding your data and no per-click meter. You provision a clean Linux VPS, install the software, point your campaigns at it, and it reads paid traffic to offers and funnels you route yourself, on a flat licence that does not climb with clicks. Read the full Keitaro review for the detail.

Measured is an enterprise media-effectiveness platform, and it reaches a related goal by the opposite route. It does not sit in your click path and it does not fire a pixel. It asks which channels are actually causing sales across Meta, Google, TikTok, connected TV, affiliates, email and offline, and which are taking credit for sales that would have happened anyway. It answers that with two methods on one system: geo-matched incrementality experiments with holdouts that measure a channel's true lift, and a media mix model calibrated by those experiments so the model is anchored to causal evidence rather than correlation alone. On top sit a Media Plan Optimizer, cross-channel reporting and peer benchmarks. Where Keitaro tracks the click on your own hardware, Measured runs experiments on the whole media mix and hands finance and marketing one planning number. The Measured review covers who it fits.

Your own server versus a serviced platform

Before either tool's method matters, there is a plainer split: who runs the thing. Keitaro is software, and you are the operator. Its own docs require a clean VPS on CentOS 9 or 10 Stream with KVM, 20 GB of SSD and 4 GB of RAM or more, and the real cost is the server, tracking domains, backups, monitoring and uptime on top of the licence. The upside is control and a fixed bill: your click and conversion data lives on your hardware with retention you set, and the licence is the same whether you push a million clicks or two hundred million. For a buyer already comfortable running infrastructure, that is the cheapest way to track at high volume. For a buyer who is not, it is a second job, and the sharpest Keitaro complaints, such as 502 errors on conversion postbacks, usually trace to an under-provisioned box rather than the tracker. See how that trade reads against a cloud tracker in Keitaro vs PeerClick.

Measured is the opposite: fully serviced, and you never touch a server. Guided onboarding, weekly model refreshes, help winning internal buy-in and peer benchmarks are part of what you pay for, because running causal experiments alone is beyond most lean marketing teams. That hand-holding is real value at enterprise scale, and it is priced accordingly. The two sit at opposite ends of the same axis: Keitaro asks you to run everything and charges a flat fee, and Measured runs everything for you on an annual contract you cannot see the price of until you book a demo.

Deterministic click tracking versus causal experiments

This is the split under everything else. Keitaro is deterministic. It stamps a click, follows it to a conversion, and reports at the level of the specific campaign, source and creative, so it can tell you which placement produced a sale and what each one truly returned against real spend. That click-level read is why a media buyer pays for it: the number follows the click. What it does not do is prove causation. A click tracker credits the touch it can see; it cannot tell you whether a sale would have happened without that channel.

Measured never follows a click. It runs a real experiment: hold a channel out in matched geographies, watch what happens to sales, and read the difference as that channel's true lift. Those lift results then calibrate a media mix model that explains total revenue as a function of spend over time. That is a fundamentally different kind of proof. It can value connected TV, podcasts, direct mail and upper-funnel prospecting, the channels a click tracker structurally cannot see, because turning a channel off and measuring the drop does not depend on a click at all. The cost is granularity. A geo test tells you a channel drove incremental sales; it does not tell you which specific ad closed a named buyer. For that, Keitaro gives the cleaner answer.

Neither method is better in the abstract. Click-level tracking is right when you route paid traffic yourself and need to price each source and creative. Causal experiments plus a calibrated model are right when your core question is how to split the next budget dollar across many channels, including offline and upper-funnel, and you want that answer proven by a holdout rather than inferred from a click.

The feedback loop: only Keitaro closes it

This is the one axis where the two are not merely different but opposite, and it is where Keitaro pulls ahead for a paid-traffic buyer. Keitaro feeds conversions back. It fires server-to-server postbacks and syncs conversions natively to Meta, TikTok and Google, so the ad platforms optimise their delivery on the sales your server actually recorded rather than only the ones their own pixel caught. For a Meta or TikTok buyer whose in-platform tracking loses conversions to iOS and cookie loss, that returned signal is a large part of the value: cleaner conversion data going back in usually means cheaper, better-targeted delivery coming out. Measured feeds nothing back at all. It is a measurement and planning layer, not a conversion-API pipe, so it sends no signal to the ad platforms and is not the tool that closes this loop.

Measured's counterweight is reach, not feedback. Because it proves lift by experiment rather than by click, it sees the parts of the funnel neither a pixel nor a click tracker can reach: prospecting social, YouTube, connected TV, podcasts and offline media, plus saturation curves that show where a channel stops paying off. Its Media Plan Optimizer turns those reads into a spend recommendation. So the split is clean. Keitaro gives you click-level depth, a fixed cost on traffic you route yourself, and a real signal wired back to the platforms; Measured gives you a causal, cross-channel view of where the next budget dollar should go, with nothing sent back.

Why your numbers will not match Meta or Shopify

Both tools disagree with your ad platforms, and they disagree in opposite ways. Keitaro is deterministic, so its counts should sit close to your ad accounts once tracking is clean, with the differences down to how it stamps and de-duplicates clicks and how its postbacks reconcile against the platforms. Measured runs the other way: an experiment and a mix model produce estimates and confidence intervals, not deterministic user counts, so they will not reconcile with Meta, GA4 or Shopify, and they will usually strip credit off channels that self-report generously. That independent read is the whole point of buying Measured. Treat whichever you choose as a decision layer, reconciled against backend revenue, and keep your attribution windows consistent when you compare.

Who each one is for

Keitaro fits the operator who buys high-volume paid traffic and routes it to offers and funnels they control, is comfortable running a server, and wants deterministic per-source tracking with conversions synced back to Meta, TikTok and Google, on a bill that does not climb with clicks. Its flat licence is what pays off once click counts are large, one box covers native, push, pop and search offers at once, and bot filtering ships in the tracker rather than as an add-on. Its ceiling is that it is a self-hosted media-buying tracker, not a causal-measurement platform: it will not prove the incremental lift of connected TV or upper-funnel prospecting, because those channels do not resolve to a click. See the alternatives to Keitaro if running a server is the sticking point.

Measured is the fit when you spend at least six figures a month across several channels, run media a click tracker cannot see such as connected TV or upper-funnel prospecting, and have finance challenging marketing's numbers, because a test-calibrated model is exactly what settles that argument. It is the wrong tool for a buyer routing paid clicks to offers they run, where a tracker such as Keitaro or Voluum does the job, and its floor and heavy onboarding rule it out below serious spend. One caveat covers both. Below roughly $50,000 to $100,000 a month of spend, clean UTMs, server-side events and blended metrics like MER and new-customer CAC do most of the job for far less, and neither tool fixes a weak offer, a broken pixel setup or a low conversion rate. They make the numbers clearer. They do not make the funnel better. For the full field, see the best ad tracking and attribution software.