Hyros and Ripplux both exist because platform-reported numbers are not trustworthy. They respond in opposite ways. Hyros tries to see more of the customer journey than the platforms do, recovering conversions it says the platforms drop. Ripplux stops trying to see the journey and runs an experiment instead.
Deciding between them comes down to a single question: do you want a more complete record of what happened, or evidence about what caused it?
What each tool actually does
Hyros describes its method as AI attribution, positioning it against conventional pixel tracking, and advertises recovery of conversions the ad platforms miss. It supports call tracking for businesses that sell through booked calls, and its site names four customer types: SaaS, e-commerce, call-based businesses, and info or education products. Agencies are called out separately. That breadth is worth noting, because it shapes the product: Hyros is not Shopify-first, and a good deal of its heritage is in long-window, high-ticket, call-driven funnels where a single conversion can take weeks and involve a human on the phone.
Ripplux is Shopify-only and does one thing. It reads your order history and ad spend through official APIs, then runs randomised holdout experiments on Meta and Google. A randomised share of your audience is withheld from a campaign, and the difference in orders between the exposed and withheld groups is the campaign's incremental effect.
Feature comparison
| Hyros | Ripplux | |
|---|---|---|
| Core method | Attribution with conversion recovery | Randomised holdout experiments |
| Tracking script required | Yes | No |
| Identity graph | Yes | No |
| Call tracking | Yes | No |
| Long-window attribution | A core selling point | Not applicable, tests measure a window directly |
| Answers "did this ad cause the sale" | No, it reports observed paths | Yes, within a confidence interval |
| Platform focus | SaaS, e-commerce, call funnels, info products | Shopify e-commerce only |
| Published pricing | No, demo required | Yes, $149 and $299 per month |
| Pricing basis | Reported to scale with tracked revenue | Flat per plan |
The privacy argument, told accurately
Comparison pages in this category, including an earlier version of this one, tend to argue that tracking tools are doomed because Chrome is about to remove third-party cookies. That argument is out of date and we are not going to make it.
Google abandoned the plan. It announced in July 2024 that it would not phase out third-party cookies in Chrome, confirmed in April 2025 that it would not ship the standalone choice prompt, and shut down most of the Privacy Sandbox initiative in October 2025. As of 2026, third-party cookies remain in Chrome.
The honest version of the privacy point is narrower and still real. Safari, Firefox, and Brave block third-party cookies today, which removes a meaningful share of traffic from cross-site tracking regardless of what Chrome does. Apple's App Tracking Transparency continues to limit in-app identifiers. Consent requirements under GDPR and a growing set of US state laws mean a portion of your visitors are never trackable at all. None of that is a cliff. It is a permanent, partial ceiling on how much of the journey any tracking product can observe.
So the case for experiments is not that tracking is about to break. It is that tracking answers a different question.
Why better tracking still cannot prove causation
Suppose Hyros works perfectly and recovers every conversion the platforms miss. You now know, with complete accuracy, which ads each buyer saw before purchasing.
That still does not tell you whether the ad caused the purchase.
The clearest case is branded search. Someone who already decided to buy from you searches your brand name, clicks the ad sitting above your own organic result, and converts. Every attribution system on the market, tracking perfectly, credits that ad. Withhold the ad and most of those buyers click the organic link and purchase anyway. The ad was on the path and was not the cause. Field experiments on branded search have found the incremental effect close to zero, which is why the eBay and Econometrica study remains the reference point for this problem.
Retargeting has the same shape. You are showing ads to people who already visited your store, which is to say people already more likely to buy. Attribution credits the retargeting campaign. An experiment tells you how much of that would have happened anyway.
No amount of tracking fidelity resolves this, because the missing information is not about what happened. It is about what would have happened otherwise, and that is unobservable by definition. The only way to estimate it is to create a comparable group that did not see the ad.
When Hyros is the better choice
If you sell through booked calls, Hyros does call tracking and Ripplux does not. That is a genuine capability gap and it is not close.
If you are not on Shopify, Ripplux does not serve you at all. Hyros covers SaaS, info products, and coaching businesses that Ripplux has no path to support.
If your purchase window is genuinely long, measured in weeks or months with multiple human touches, per-journey attribution gives you a picture of that sequence that an aggregate experiment does not.
And if your monthly ad spend is below roughly $5,000, Ripplux cannot run a valid holdout for you and will say so. A tracking product will still produce output at that scale, and some visibility beats none.
When Ripplux is the better choice
You are a Shopify merchant spending somewhere between $5,000 and $50,000 a month on Meta and Google, and you suspect that a chunk of your reported ROAS is credit for sales that would have happened anyway.
You do not want a tracking script in your theme, either because of consent-mode constraints or because you would rather not add another dependency to the checkout path.
You want a number you can defend. An experiment result comes with a confidence interval and a method you can describe to anyone who asks. An attribution figure comes with a model.
What Ripplux will refuse to tell you
Holdout experiments need volume to produce a usable answer. Ripplux requires roughly $5,000 in monthly spend on the platform being tested and about 50 conversions per month. Below that, the confidence interval is too wide to support a decision, and Ripplux reports that rather than a number that looks precise and is not.
Ripplux also does not decompose a campaign's effect by touch type, does not measure view-through separately, and does not track anything at the level of an individual customer. Those are deliberate omissions, and if you need them, this is the wrong tool.
Can you run both?
They stack more awkwardly than most pairs, because they disagree by construction. Hyros will report more conversions than the platforms do. A holdout will often report fewer incremental orders than any attribution model. When you see both numbers, the temptation is to average them, and that produces a figure with no meaning.
The workable arrangement is to treat them as answering separate questions and never to reconcile them. Use attribution for pacing and diagnostics day to day. Use an experiment when you are deciding whether a whole campaign or channel deserves its budget. If you find yourself trying to make the two numbers agree, you are misusing at least one of them.
Frequently asked
- Does Ripplux track customers across sessions?
- No. Ripplux maintains no user identity graph and installs no tracking script. It reads Shopify orders and ad spend through official APIs and runs randomised holdout experiments on the ad platforms. The causal estimate comes from comparing two randomised groups, not from following individuals.
- Is Ripplux affected by iOS App Tracking Transparency or browser cookie blocking?
- A holdout experiment is not, because it does not rely on identifying who saw what. Randomisation happens on the ad platform and the outcome is measured in aggregate order counts. Tracking-based tools lose signal as consent rates and cross-site identifiers shrink, which affects coverage rather than randomisation.
- Is better tracking the same as knowing what worked?
- No, and this is the central difference. Perfect tracking tells you which ads a buyer saw. It cannot tell you whether that buyer would have purchased anyway. Only withholding the ad from a comparable randomised group answers that, because it is the only method that changes something and observes the result.
- Does Hyros publish its pricing?
- As of 2026-07-28, Hyros does not publish list pricing on its site and routes prospects to a demo. Third-party summaries report tiers based on tracked revenue, starting around $69 per month for a small Shopify tier and rising into four figures at higher tracked-revenue bands. Treat those figures as reported rather than confirmed, and get a quote.