Google Ads Bidding Strategies: Maximize Conversions & Target CPA Mastery

The conversation about Google Ads bidding strategies in 2026 goes like this: pick Maximize Conversions to gather data, graduate to Target CPA once you have 30 to 50 clean conversions per month, layer in Target ROAS when your values are consistent. Repeat the cycle. Everyone who runs paid search knows this playbook. What nobody is saying out loud is that the playbook assumes your conversion data is real — and in 2026, a meaningful percentage of it is not.

That gap is where most advertisers are actually losing money. Not from choosing the wrong bidding strategy. From feeding a machine learning system the wrong inputs and then optimizing the output like the math makes sense.

This guide covers every major bidding strategy, every tool worth using to support it, and the one layer the strategy guides conveniently skip.

What changed in 2026 that breaks the old mental model

In January 2026, Google launched Google Tag Gateway for free. One Cloudflare worker, one subdomain, Google tags load first-party. That eliminated the signal-loss argument for Google-only advertisers at the SMB level. In April 2026, Meta launched free one-click CAPI. If you run a single Shopify store advertising only on Meta, your conversion tracking cost just dropped to zero.

Both moves commoditized the pipe. The pipe is not the problem anymore.

Then ChatGPT Ads Manager launched on May 5, 2026, with its own CAPI endpoint. Seventy point six percent of LLM-referred traffic currently misclassifies as direct in GA4. If you are running any form of performance creative through AI-driven channels and your attribution model depends on GA4 last-click, your CPA numbers are wrong before you choose any bidding strategy at all.

Project Andromeda, fully deployed by Google in October 2025, acts on contaminated signals within hours. That means bot-sourced conversions flowing into your Smart Bidding campaigns do not sit around waiting for an audit. The algorithm learns from them immediately and starts bidding accordingly in the next auction cycle.

The bidding strategy debate — Max Conversions versus Target CPA versus Target ROAS — is real and worth having. But it is downstream of a more fundamental question: what is in your conversion pipeline right now, and how much of it is real?

Quick answers

What is the difference between Maximize Conversions and Target CPA?

Maximize Conversions tells Google to spend your entire daily budget getting as many conversion events as possible, with no constraint on cost per conversion. Target CPA tells Google to spend whatever budget is available while keeping the average cost per conversion at or below your specified target. Setting a Target CPA inside a Maximize Conversions campaign produces behavior identical to the standalone Target CPA strategy. The algorithm is the same either way.

How many conversions do you need before Smart Bidding works?

Google's documented minimum is 15 conversions in 30 days to enter Smart Bidding at all. In practice, Target CPA needs 30 to 50 monthly conversions to produce stable performance. Target ROAS needs 50 or more, because the algorithm is optimizing for revenue value, not just event count. Below these thresholds, the learning phase extends indefinitely and campaign performance is erratic. The problem nobody states: if 20 to 30 percent of those conversions are bot-generated form fills or ghost events, you need proportionally more real conversions to hit the effective threshold. Your dashboard says you have 50. Your algorithm may be training on 35.

What is the Smart Bidding learning phase?

After any significant campaign change — new bid strategy, budget shift above 20 percent, new conversion action — Google requires roughly 50 conversion events or three full conversion cycles to recalibrate the algorithm. During this window, performance is unstable and costs typically run higher. The 50-conversion guideline remains accurate in 2026, with one nuance: signal consistency matters as much as volume. Fifty conversions arriving in irregular bursts with wildly varying values may extend the learning period further than 50 conversions distributed evenly across the month.

Should you use portfolio bid strategies?

Portfolio strategies let you pool conversion data across multiple campaigns, which solves the volume problem for accounts where individual campaigns cannot hit the 30 to 50 monthly conversion threshold on their own. Google recommends against hard bid caps inside portfolio strategies because they constrain the algorithm from entering high-value auctions. The tradeoff is real: caps protect CPA in the short term, caps limit reach in auctions where conversions are most likely. Start without caps, set them only after you have observed the algorithm's natural CPA range across at least four weeks.

When should you switch from Maximize Conversions to Target CPA?

Start with Maximize Conversions for two to four weeks to let the algorithm accumulate auction-time signal data. Once you consistently hit 30 or more conversions per month and your observed CPA has stabilized, introduce a Target CPA at approximately 80 percent of your maximum acceptable CPA. This gives the algorithm room to optimize without immediately restricting the auctions it can enter. If performance deteriorates after switching, return to Maximize Conversions briefly to let the algorithm re-stabilize before reintroducing a target.

Does manual CPC still work in 2026?

Manual CPC is not dead, but Google functionally disadvantages it. Accounts relying on manual bidding see reduced impression share as Google's auction-time signals favor campaigns running Smart Bidding. Manual CPC is defensible for brand campaigns with predictable CPC ranges and for small budgets where conversion volume is too low for Smart Bidding to function at all. Everywhere else, the gap between manual bidding and a well-fed Smart Bidding strategy has become too wide to defend on performance grounds alone.

What is Google Tag Gateway and does it replace server-side GTM?

Google Tag Gateway is a free reverse proxy that routes GA4 and Google Ads tags through your own subdomain instead of Google's CDN. It bypasses ad blockers for Google-specific tags and improves script load performance. It does not replace server-side GTM for multi-platform setups. GTG cannot route Meta CAPI, TikTok Events API, or LinkedIn Conversions API events. It does not transform or filter data in transit. If you advertise exclusively on Google, GTG is sufficient and costs nothing. If you run any other paid platform, you need a server-side layer alongside it.

The real problem nobody names in bidding strategy guides

Smart Bidding is a closed-loop learning system. Every conversion event you send becomes a training signal. The algorithm uses that signal to model which auctions, which audiences, and which creative combinations are most likely to produce more conversions like the ones you already sent.

This architecture works brilliantly when conversions are real. It destroys budget quietly when they are not.

Form spam fires your Google Ads conversion tag before any validation happens. Smart Bidding logs the event as a successful conversion. The algorithm updates its model: this type of traffic, this auction, this time of day produced a conversion. It bids higher on similar traffic in the next cycle. That traffic produces more form spam. The loop closes.

Bad bots — scrapers, credential stuffers, automated form submitters — represented 37 percent of all web traffic in 2024 and reached 45 percent of US internet traffic by early 2026 (Cloudflare). SpiderAF's analysis of 4.15 billion clicks found a click-fraud rate of 5.12 percent on average, with extreme cases reaching 51.8 percent of ad budget lost to bots. Global invalid traffic runs at 20.64 percent (Fraudlogix 2026). Google's own Audience Network sits at 67 percent IVT.

The PillarlabAI case is worth reading once if you run lead generation. Four thousand five hundred sixty signups in four weeks. Seven hundred thirty real humans. Eighty-four percent fraudulent. Six hundred fifty accounts from a single laptop. Every one of those fake signups fired a conversion event. Every one trained the bidding algorithm to find more traffic like the bot.

That is the actual bidding problem in 2026. Not which dial to turn. What is flowing through the pipe you are dialing.

Bidding strategies explained

Maximize Conversions

The starting point for most campaigns with limited history. Google spends your full daily budget to collect as many conversion events as possible, with no constraint on CPA. This strategy is appropriate when you do not yet know what your sustainable CPA looks like, when you have fewer than 30 monthly conversions and cannot yet graduate to Target CPA, or when you are re-entering a learning phase after a major campaign change.

The structural risk: Maximize Conversions will always spend your full daily budget. There is no efficiency brake. If your conversion tracking is broken, duplicate-tagged, or contaminated with bot events, the algorithm optimizes aggressively toward whatever it is receiving. Budget burns. Reported conversions climb. Real business outcomes stay flat or decline. This is not a theoretical risk. It is the most common pattern in accounts where tracking has not been audited recently.

Standalone campaigns using Maximize Conversions must have their own individual daily budgets. Shared budgets cause unpredictable pacing behavior.

Target CPA

The mature version of Maximize Conversions. You set a target cost per conversion. The algorithm adjusts bids auction by auction to achieve that target on average across the campaign. It does not guarantee every conversion costs exactly your target — some will cost more, some less — but the campaign average should approach your target over time.

Minimum effective threshold: 30 conversions per month at the campaign level. Below that, performance is erratic. The practical floor that produces visibly stable campaigns is closer to 50 per month. If individual campaigns cannot reach this threshold, pool them into a portfolio bid strategy to hit it collectively.

Setting your initial target: use your observed average CPA from the Maximize Conversions phase as a baseline. Start at or near that number, not at your theoretical maximum acceptable CPA. If your observed CPA is $60 and you set a target of $30, the algorithm is mathematically restricted from entering most of the auctions where it previously found conversions. Traffic drops. The campaign appears to underperform and gets adjusted again, resetting the learning phase.

The common error: setting an aspirational target CPA instead of a realistic one, watching performance drop, concluding Target CPA does not work, returning to Maximize Conversions, and repeating the cycle indefinitely. The strategy is not broken. The target was wrong.

Target ROAS

Target ROAS optimizes for conversion value rather than conversion volume. You tell Google what return you want on every dollar of ad spend. The algorithm bids higher on clicks predicted to generate more revenue and lower on clicks predicted to generate less, regardless of how many conversions that produces.

This strategy requires conversion value tracking — you must be passing actual revenue data with every conversion event, not just a fixed placeholder value. It requires 50 or more monthly conversions with value data attached. It is the correct strategy for e-commerce operations with variable order values, subscription services with different plan tiers, and B2B accounts where deal size varies materially.

The EMQ factor matters here. When your conversion API event match quality (EMQ) score moves from 8.6 to 9.3 — meaning your server-side events are carrying clean, matched customer data — Meta advertisers see 18 percent lower CPA and 22 percent ROAS lift on average. Google's Enhanced Conversions work on an equivalent principle. Better signal in, better auction targeting out. The bidding strategy choice is secondary to the quality of the signal feeding it.

Maximize Conversion Value

The budget-spending counterpart to Target ROAS. The algorithm chases the highest-value conversions it can find without a return constraint, spending your full daily budget to do it. Appropriate for campaigns with new budget or where you need to establish what your achievable ROAS range looks like before setting a constraint. Carries the same risks as Maximize Conversions: full budget spend regardless of efficiency, aggressive optimization toward whatever conversion events it is receiving.

Portfolio bid strategies

Portfolio strategies pool conversion data across multiple campaigns under a single bid strategy. This solves the data-volume problem for accounts where individual campaigns cannot reach the 30 to 50 monthly conversion threshold on their own. A portfolio of five campaigns averaging 10 conversions per month each gives the algorithm 50 events monthly to work with — functionally equivalent to one well-fed campaign.

Google recommends against bid caps inside portfolio strategies. The argument: caps prevent the algorithm from entering high-value auctions precisely when conversion probability is highest, which defeats the purpose of Smart Bidding. The counterargument from practitioners: uncapped strategies in contaminated accounts will bid aggressively on bot traffic at full CPA or higher. The right answer depends entirely on the quality of your conversion data. Clean data: remove caps and let the algorithm work. Contaminated data: fix the contamination first, then remove caps.

Performance Max

PMax is not a bidding strategy in the traditional sense. It is a campaign type that absorbs your bidding strategy and applies it across all Google inventory simultaneously: Search, Shopping, Display, YouTube, Gmail, Maps, Discover. The bidding signals you feed it — conversion events, customer data, creative assets — determine where it allocates budget automatically.

PMax amplifies whatever is in your conversion pipeline. High-quality signals with clean bot-filtered event data: the algorithm finds real buyers across channels efficiently. Contaminated signals: the algorithm finds ways to spend your budget across the entire Google network on traffic that looks like what you sent it. There is no manual override for where PMax spends. The only lever you have is the quality of the signal.

The tool landscape

Every tool in this category falls into one of two jobs: delivering conversion events to Google's bidding infrastructure (the pipe), or ensuring those events represent real human actions (the water). Most tools focus exclusively on the pipe. A small number address both. That distinction is what the comparison below is built around.

DataCops

DataCops is the only tool in this category that combines bot-filtered first-party CAPI, a first-party CMP, and cookieless persistent identity resolution in a single architecture. The distinction matters for bidding strategy performance: if your Smart Bidding campaign is training on conversion events, the quality of those events directly determines the quality of the model. DataCops filters bot traffic against a 361B-plus IP database before any conversion event fires. Events that reach Google's bidding infrastructure represent verified human actions.

The first-party CMP loads from your own subdomain rather than a third-party CDN. Competitor CMPs (OneTrust, Cookiebot) load from third-party CDNs that uBlock Origin and Brave block 30 to 40 percent of the time. When the CMP does not load, the consent gate does not function, and you lose the ability to distinguish consented identifiable data from anonymous data. DataCops' consent gate loads on every session because it is not on any filter list.

The cookieless persistent identity architecture re-identifies returning users without relying on cookies. No ITP degradation. No seven-day expiry. EU users see the TCF 2.2 consent banner and identity resolution activates upon consent. Non-EU users get persistent identity by default, because no legal requirement exists. Every competitor either uses cookies (killed by ITP in seven days) or goes fully cookieless (loses returning users entirely). Neither approach produces complete funnel data for Smart Bidding to train on.

Setup: one script tag plus one CNAME record. Live in five to 30 minutes. No developer needed. Works on Shopify, WooCommerce, Webflow, and custom stacks.

CAPI platforms: Meta, Google Enhanced Conversions, TikTok Events API, LinkedIn Insight. No Pinterest. No Snapchat.

What does not work: SOC 2 Type II is in progress, not yet certified. Integration catalog is narrower than Tealium or Segment for enterprises needing deep custom CRM pipelines. Not the right call for enterprises with dedicated tagging engineering teams who want full container-level control.

Right for: Any advertiser running Smart Bidding who wants to verify their conversion events represent real humans before they feed the algorithm. E-commerce and lead gen operations on Google plus at least one other paid platform. Accounts where bot contamination in the conversion pipeline has been creating unexplained CPA volatility.

Value: 9/10. The only tool bundling bot filtering with CAPI at SMB pricing.

Pricing: Free (2,000 sessions, no CAPI). Growth $7.99/month (5,000 sessions, no CAPI). Business $49/month (50,000 sessions, full CAPI across all platforms). Organization $299/month (300,000 sessions). Enterprise custom.

Google Tag Gateway

Google Tag Gateway is Google's own free reverse proxy for Google tags. Your gtag.js and GA4 collection requests route through a subdomain you control rather than Google's CDN. Ad blockers cannot identify the requests by hostname. Scripts load faster. First-party context improves.

What it does not do: GTG does not write server-side cookies. ITP's seven-day cap on JavaScript-written cookies still applies. GTG does not filter or transform data in transit. GTG routes only Google tags — GA4, Google Ads, Floodlight. Meta CAPI, TikTok, LinkedIn: none of these are covered. If your entire paid stack is Google-only and you do not need multi-platform CAPI or data enrichment, GTG is the correct starting point and costs nothing.

Right for: Google-only advertisers at any spend level wanting basic ad-blocker bypass without infrastructure costs.

Value: 10/10 for the use case it covers. Free with no meaningful setup friction.

Pricing: Free.

Stape

Stape is the most widely used server-side GTM hosting platform. It runs your server container on Google Cloud Run with managed infrastructure, removing the DevOps requirement from sGTM deployment. The Stape Gateway product mirrors GTG functionality for GA4 and Google Ads through a simpler interface. The broader sGTM hosting enables the full multi-platform CAPI stack: Meta, TikTok, LinkedIn, Pinterest, Snapchat, and 80-plus GTM template integrations.

What Stape does not do: Stape is infrastructure, not a product. You still need GTM expertise to configure the container, set up event matching, and maintain templates as vendor specs change. There is no bot filtering. Invalid traffic that reaches your tags gets forwarded to every connected CAPI endpoint unchanged. For Smart Bidding purposes, Stape delivers the pipe but makes no claim about the water quality flowing through it.

Bounteous research found that 80 percent of sGTM setups are detectable by sophisticated ad blockers due to consistent configuration patterns. Stape's CNAME-based implementation reduces this risk but does not eliminate it.

Right for: In-house marketing engineers or agencies with GTM expertise who want maximum flexibility and template coverage without managing cloud infrastructure directly.

Value: 7/10. Excellent infrastructure for teams with the technical capacity to use it. Assembly required.

Pricing: $17/month Pro. $83/month Business. Cloud Run costs $50 to $300/month additional depending on traffic volume.

Elevar

Elevar is the dominant Shopify-native CAPI solution for high-volume e-commerce. It integrates directly with Shopify's order data layer, enabling millisecond-level order event tracking with full order-level fidelity. The session enrichment layer appends user identity data to server-side events, improving EMQ scores. Elevar has deep Shopify expertise and handles the January 13, 2026 App Pixel default change to "Optimized" — which silently throttled client-side pixels with no notification from Shopify — better than most alternatives because it was already predominantly server-side.

What Elevar does not do: Elevar is Shopify-only. If you run any non-Shopify storefront or B2B SaaS alongside your DTC brand, Elevar does not cover it. Pricing scales sharply with order volume — $200 per month at 1,000 monthly orders is defensible; $950 per month at 50,000 orders is a substantial CAPI line item for a mid-market brand. There is no bot filtering. Invalid order events and fraudulent signups flow to Meta CAPI and Google Enhanced Conversions alongside legitimate ones.

Right for: Shopify-only seven-figure stores where order-level fidelity and tight Shopify data layer integration are worth the premium.

Value: 7/10. Best-in-class for its specific use case. Expensive outside the Shopify context.

Pricing: $200/month Essentials (1,000 orders). $950/month Business (50,000 orders).

Tracklution

Tracklution is a European-market-oriented server-side tracking platform with TCF 2.2 consent compliance built in. It covers Meta CAPI, Google Enhanced Conversions, TikTok Events API, and several other platforms. Setup is simpler than raw sGTM and does not require GTM expertise. SOC 2 Type II and ISO 27001 certified, which matters for EU enterprise buyers with procurement requirements.

What Tracklution does not do: no bot filtering. Events are delivered server-side but no validation of whether the conversion originated from a human versus automated traffic. For EU-focused agencies managing multiple clients with GDPR compliance requirements and limited technical resources, Tracklution is a strong operational fit. For advertisers concerned about bot contamination in their Smart Bidding signal, Tracklution solves the pipe but not the water.

Right for: EU-focused small to mid-market agencies wanting SOC 2 certified server-side tracking without GTM overhead.

Value: 7/10. Clean EU compliance story, honest pricing.

Pricing: €31/month Starter. Custom Enterprise.

Aimerce

Aimerce targets mid-market e-commerce with a server-side CAPI stack focused on identity resolution and session recovery. The identity matching layer attempts to re-identify users who have dropped cookies or switched devices, improving event match quality for Meta and Google. The product is positioned between Elevar's Shopify depth and Stape's infrastructure flexibility.

What Aimerce does not do: no bot filtering. Pricing scales usage-based above 1,000 orders per month, which creates unpredictable cost structures for fast-growing accounts. Less template depth than Stape for non-standard integrations.

Right for: Mid-market DTC brands on Shopify or WooCommerce running $500K to $5M GMV annually where session recovery and identity matching justify the cost.

Value: 6/10. Solid middle-market product. Pricing model can be difficult to forecast.

Pricing: $299/month base. Usage-based above 1,000 orders.

Littledata

Littledata is a Shopify and headless commerce-focused data pipeline that specializes in connecting server-side order data to GA4 and Meta. It handles the Shopify-to-GA4 data quality problem that client-side tracking consistently fails on: purchase events firing before payment confirmation, checkout abandonment data gaps, and subscription revenue tracking for Recharge and Bold users.

What Littledata does not do: no bot filtering. Limited platform coverage compared to full sGTM setups. Less flexible for custom event schemas outside the core e-commerce purchase funnel.

Right for: Shopify brands where GA4 data accuracy for Smart Bidding training is the primary concern and the existing stack does not include subscription or headless commerce complications.

Value: 6/10. Solves a specific problem well. Expensive relative to the scope of what it covers.

Pricing: $89/month and scales per order volume.

TrackBee

TrackBee is a Dutch-market-origin server-side tracking tool with a straightforward Meta CAPI and Google Enhanced Conversions implementation. Clean UI, reasonable European pricing, cookie recovery tooling that attempts to match anonymous sessions to known users post-consent. Gaining traction among European agencies wanting a managed server-side product without Elevar's Shopify pricing structure.

What TrackBee does not do: no bot filtering. Smaller integration catalog than Stape. Less documented than Elevar for complex Shopify data layer setups.

Right for: European mid-market e-commerce brands running Meta plus Google who want managed server-side without GTM complexity.

Value: 6/10. Reasonable option in its segment.

Pricing: €79/month and up.

Datahash

Datahash is an enterprise-tier first-party data platform focused on hashed audience matching and server-side event delivery for large advertisers. It integrates with CDP and CRM systems to enrich conversion events with first-party identity before sending to Meta CAPI, Google, TikTok, and LinkedIn. The use case is less about fixing broken tracking and more about improving event match quality for advertisers who already have mature data infrastructure.

What Datahash does not do: no bot filtering. Pricing is enterprise custom — most deployments run $500 to $2,000 per month minimum and require a sales process. Not appropriate for self-serve SMB use.

Right for: Enterprise advertisers with existing CDP or CRM infrastructure looking to improve EMQ scores across multiple platforms.

Value: 7/10 for its target buyer. Wrong tool for anything under $500K annual ad spend.

Pricing: Custom quote. Estimated $500 to $2,000/month range.

Addingwell (now Didomi)

Addingwell was acquired by Didomi in April 2025 for $83 million, combining EU consent management infrastructure with server-side tagging. The combined product is one of the clearest signals of where the market is going: CMP and CAPI converging in a single vendor, especially for EU advertisers navigating Google Consent Mode v2 (mandatory EEA from June 15, 2026). The free tier covers 100,000 requests per month, making it competitive for smaller publishers.

What Addingwell/Didomi does not do: no bot filtering. Complexity increases significantly for non-GTM implementations. US-market pricing and support are less established than EU counterparts.

Right for: EU-focused publishers and mid-market brands where Consent Mode v2 compliance and server-side tagging need to come from the same vendor.

Value: 7/10 for EU market. Pricing and support outside EU still developing.

Pricing: Free to 100,000 requests/month. Paid tiers EUR-based.

Meta 1-Click CAPI

Meta's native CAPI implementation launched April 15, 2026, and reset the floor price for Meta-only server-side tracking to zero. One click in Business Manager, integration complete. No code, no infrastructure. For single-store advertisers whose entire paid stack is Meta, there is now no rational argument for paying a third party to replicate what Meta provides natively.

What Meta 1-Click CAPI does not do: covers only Meta. No Google Enhanced Conversions, no TikTok, no LinkedIn. No bot filtering — invalid traffic reaches Meta CAPI and trains Lookalike Audiences the same way it always did. EMQ optimization is basic; advanced identity matching and data enrichment require a third-party solution. For advertisers running multiple paid platforms, the free CAPI covers one of them and leaves the rest unaddressed.

Right for: Single-store Shopify or WooCommerce operators running Meta ads only who want functional CAPI without any setup overhead.

Value: 10/10 for the use case it covers. Unbeatable at the price.

Pricing: Free.

Triple Whale

Triple Whale is a post-purchase attribution and analytics dashboard. It is not a CAPI tool in the technical sense — it reads conversion data to build attribution models, it does not write server-side events to ad platforms as its primary function. It belongs in this conversation because its dashboards often become the source of truth for bidding strategy decisions at DTC brands, and those dashboards inherit whatever corruption exists in the upstream conversion pipeline.

Triple Whale's Pixel and server-side tracking are CAPI-adjacent features. The core product is attribution modeling and cohort analysis. If your CAPI stack is clean, Triple Whale's attribution is valuable. If your CAPI stack is contaminated with bot events, Triple Whale charts those contaminated numbers beautifully. The problem is upstream of any dashboard. See advanced conversion tracking implementation for what fixing the upstream looks like.

Right for: DTC brands above $1M annual revenue wanting multi-touch attribution modeling alongside operational metrics (AOV, LTV, cohort repeat rate). Not a replacement for a server-side CAPI stack.

Value: 6/10 for its defined use case. Category mismatch if you are shopping for CAPI delivery.

Pricing: $179/month annual. $259/month Advanced. GMV-based pricing above $5M.

Northbeam

Northbeam is enterprise-tier multi-touch attribution with machine learning-modeled attribution across paid channels. It operates at a level above the CAPI conversation — it is reading what CAPI pipelines deliver and applying modeling to understand incrementality, channel interaction effects, and true media mix. For brands spending $500K or more monthly on paid media, Northbeam provides analytical depth that no CAPI tool offers.

What Northbeam does not do: it does not deliver server-side conversion events to ad platforms. It does not filter bots. It is reading contaminated data from the same upstream sources that contaminate Smart Bidding, just with more sophisticated modeling applied to it. Garbage in still produces modeled garbage out, with better charts.

Right for: Brands at $5M or more annual ad spend where channel-level incrementality measurement and media mix modeling are the actual analytical question.

Value: 7/10 for its target buyer. Completely wrong category for advertisers wanting CAPI implementation.

Pricing: $1,500/month entry. Scales to $5,000 to $10,000 or more for large accounts.

Hyros

Hyros positions itself as a phone-call and high-ticket lead generation attribution tool with its own first-party tracking layer. It focuses on the specific pain point of tracking lead-to-revenue attribution for businesses where the sale happens offline or on a call, not at checkout. The first-party script tracks session data without cookies in most configurations. The attribution model attempts to close the loop between ad click and closed revenue, not just form fill.

What Hyros does not do: no bot filtering of incoming lead attribution. For lead generation accounts where form spam is contaminating Smart Bidding, Hyros helps you see which leads closed but does not prevent fraudulent form fills from training the algorithm in the first place. Pricing is sales-led and opaque.

Right for: High-ticket service businesses and agencies where offline revenue attribution is the primary attribution problem, not CAPI delivery infrastructure.

Value: 6/10. Specific problem solver for a specific buyer profile.

Pricing: $1,000 to $5,000/month, sales-led.

Cometly

Cometly is a growth-stage attribution tool combining server-side pixel, multi-touch attribution modeling, and ad spend integration for Meta, Google, and TikTok. It competes with Triple Whale and Northbeam at a lower price point, targeting seven-figure DTC brands that cannot justify Northbeam's pricing but want more than GA4's last-click model.

What Cometly does not do: no bot filtering. Attribution modeling inherits upstream data quality problems. Less template depth than Stape for non-standard integrations.

Right for: DTC brands spending $100K to $1M annually on paid media who want multi-touch attribution without enterprise pricing.

Value: 6/10. Reasonable mid-market option.

Pricing: $199 to $499/month, sales-led pricing above entry tier.

SignalBridge

SignalBridge is the only other tool at SMB pricing that includes bot filtering as part of its CAPI stack. It covers Meta CAPI and Google Enhanced Conversions with IP-based bot detection running before events fire. The bot filtering is less comprehensive than DataCops' 361B-IP database approach — the vendor does not publish full database specs — but it addresses the category problem that most CAPI tools ignore entirely.

What SignalBridge does not do: no CMP included. Narrower platform coverage than full sGTM setups. Less documentation depth than established players. Smaller customer base means fewer real-world edge cases resolved in support documentation.

Right for: Small to mid-market advertisers who want basic bot filtering alongside CAPI without enterprise pricing and are comfortable with a less-established vendor.

Value: 7/10. The bot filtering inclusion at this price point is meaningful.

Pricing: $29/month.

Raw Server-Side GTM (DIY)

Building and maintaining your own sGTM container on Google Cloud Run gives you the most flexibility available in the tracking category. Full control over every event that fires, every data transformation applied before it reaches a CAPI endpoint, every consent-gating decision. The template library covers 80-plus integrations. There is no vendor dependency beyond Google's container infrastructure.

What raw sGTM does not do: it does not provide bot filtering natively. You would need to build custom validation logic or integrate a third-party IP intelligence feed into your container — which is possible but requires engineering investment. Setup cost ranges from $5,000 to $10,000 for a proper implementation. Cloud Run runs $50 to $150 per month for typical SMB traffic. Ongoing maintenance: every vendor CAPI spec update requires template updates, which require GTM expertise to implement. First-year total cost of ownership: $11,880 to $36,600 depending on Cloud Run tier and implementation fees.

Right for: Enterprises with dedicated marketing engineering teams who want full container control, custom data transformation, and no vendor dependency at scale.

Value: 8/10 for the right buyer. Wrong call for anyone without in-house GTM engineering.

Pricing: Setup $5,000 to $10,000. Cloud Run $50 to $150/month ongoing.

Buyer decision framework

Shopify brand, under $500K GMV, Google and Meta only

Start with Google Tag Gateway for Google tracking — free, 15-minute setup. Use Meta's free one-click CAPI for Meta. You are now at zero CAPI cost. Your remaining problem is bot filtering, which neither free tool addresses. If your lead or purchase quality is clean and you have no form spam issues, this stack is sufficient. If you are seeing CPA volatility or Smart Bidding campaigns that never stabilize, the contamination question is worth investigating before you adjust any bidding targets.

Shopify brand, $500K to $5M GMV, Google plus Meta plus at least one other platform

Elevar if Shopify-only and order-level fidelity is the primary concern. DataCops at Business ($49/month) if you need multi-platform CAPI, bot filtering, or a CMP included. The $49 versus $200 entry point is a real consideration — Elevar's depth at the Shopify data layer is better; DataCops' breadth across platforms and the bot filtering layer are better.

Multi-platform B2B SaaS, lead generation

This is where bot contamination in Smart Bidding does the most damage. Form spam fires conversion tags. Smart Bidding trains on form spam. The algorithm finds more traffic that looks like the bots. Lead volume climbs, lead quality falls, sales team conversion rates drop, no one connects the ad account to the CRM outcome.

DataCops' HubSpot AI lead scoring integration and SignUp Cops address the fake signup detection problem at the source. Server-side validation before the conversion event fires: clean signal in, clean algorithm training out. For pure CAPI infrastructure without bot filtering, Tracklution or Stape are the right operational choices.

EU-focused operation with Consent Mode v2 requirements

Consent Mode v2 became mandatory for EEA advertisers on June 15, 2026. Without a CMP that functions correctly and passes consent signals to Google's bidding infrastructure, your Smart Bidding campaigns are running on degraded data in the EU. The problem no one names: if your CMP loads from a third-party CDN — OneTrust, Cookiebot, Usercentrics — uBlock Origin and Brave block it 30 to 40 percent of the time. The banner does not appear. No consent is recorded. Google receives no consent signal for those sessions. Addingwell/Didomi, DataCops, or Stape combined with a compliant CMP solve this. OneTrust standalone does not, if the script is blocked before it renders.

Enterprise with $250K-plus monthly ad spend

Raw sGTM on Cloud Run, with Datahash or a CDP for identity enrichment. The data engineering flexibility justifies the setup cost at this scale. Add Northbeam or a media mix model for incrementality measurement. The CAPI pipe is table stakes; the analytical question at this budget level is which channels are genuinely incremental and how budget should shift accordingly.

Feature comparison

DataCops is the only tool in this table combining bot filtering with a built-in first-party CMP and four-platform CAPI at SMB pricing.

When NOT to use DataCops

You need SOC 2 Type II certification today. Tracklution and Datahash are certified. DataCops' certification is in progress. If your procurement process requires it before signing, DataCops is not the right call right now.

You are Shopify-only at high order volume and order-level fidelity is the core requirement. Elevar's Shopify data layer integration is deeper. The millisecond purchase event fidelity and native Shopify checkout integration that Elevar provides is a real operational advantage for seven-figure Shopify stores. DataCops is broader; Elevar is deeper on this specific platform.

You have in-house GTM engineers and want full container control. Stape plus raw sGTM gives your team maximum flexibility, a larger template library, and no vendor lock-in at the CAPI delivery layer. DataCops is a managed product. If you have the engineering capacity and the desire for full ownership, the infrastructure route wins on control.

Your entire paid stack is Google-only and you have no bot contamination concern. Google Tag Gateway at zero cost covers your tracking needs. Adding DataCops at $49/month for features you do not use is not a good value decision. Start with GTG. Revisit if you add Meta, TikTok, or LinkedIn spend later, or if Smart Bidding performance suggests a data quality problem.

The question your bidding strategy guide is not asking

Every guide in this category — including this one to some extent — starts from the assumption that your conversion data is roughly correct. Pick the right strategy for your volume, set a realistic target, let the algorithm learn. Good advice, as far as it goes.

What the guides do not tell you to do is audit what the algorithm is actually learning from.

Your Smart Bidding campaign right now has a model. That model was trained on every conversion event you sent over the past 30 to 90 days. Some percentage of those events came from bots, form spam, ghost GA4 traffic, or duplicate tag fires. The algorithm does not know the difference. It optimized toward those events exactly as hard as it optimized toward your real customers.

The practical question: if you pulled your last 30 days of conversion events and had a way to validate each one against a real human action — a real email address, a real IP, a real session — what percentage would pass?

If you cannot answer that with a number, your bidding strategy is secondary. The algorithm is already teaching itself something. The question is whether it is teaching itself to find your customers, or to find more of whatever filled your conversion log last month.