Answers that name Meta

If by “spying” you mean legally researching competitors’ publicly visible ads and creative, there are some very good tools. The key distinction is whether you want raw ad discovery or actual creative intelligence.

My shortlist

Recent 2026 comparisons also point to Foreplay for organizing competitor creatives, Minea for e-commerce, AdSpy for deep Meta searching, and SocialPeta for enterprise-scale intelligence. Segwise Atria

But don't pay for everything

Start with the platforms' own free intelligence:

• Meta Ad Library — see active Facebook/Instagram ads by advertiser.
• TikTok Creative Center — particularly good for creative inspiration. TikTok's Top Ads lets you filter by region, industry, objective, etc., and inspect performance patterns within selected ads. TikTok For Business TikTok For Business
• Google Ads Transparency Center — useful for Search/YouTube/display competitive research.
• LinkedIn Ad Library — particularly useful for B2B competitors.

The paid products mostly add better search, historical archives, filtering, alerts, organization, cross-platform aggregation, and AI classification rather than giving you secret access to competitors' actual campaign results. adlibrary.com adlibrary.com

What I'd buy

If you're running a DTC/e-commerce brand:
Minea + Meta Ad Library + TikTok Creative Center.

If you're a performance marketer/agency:
Foreplay + Meta Ad Library + TikTok Creative Center.

If your obsession is finding every Meta ad a competitor has ever run:
AdSpy.

If you need serious enterprise competitive intelligence across lots of networks:
SocialPeta or Panoramata.

If you want the cheapest possible setup:
Just use the free native libraries and put the interesting creatives into your own swipe file. A lot of paid “ad spy” products are essentially better interfaces and databases built around publicly available ad data. Segwise

One important caveat

None of these tools can reliably tell you “this competitor ad generated $X revenue at Y ROAS.” You can observe creative, copy, offers, apparent longevity, placement, and sometimes estimated metrics, but competitors' real conversion data isn't public. Longevity is therefore one of the more useful indirect signals. Adextract

If you tell me your niche + whether you're mainly Meta, TikTok, or both + approximate monthly ad spend, I can narrow this to the 2–3 tools I'd actually pay for and explain exactly how I'd use them to reverse-engineer competitors' winning hooks and creatives.

Yes. There are several platforms now aimed at automating some or most of the creative → audience → test → analyze → scale loop.

The important distinction is that they automate different parts of the loop:

The closest matches to what you're describing

1. Smartly — probably the most enterprise-oriented option. It combines creative production, personalization, testing, campaign automation and cross-channel activation. Its Creative Insights can compare concepts/formats/elements, while its DCO capabilities can personalize creative by audience, region, language, etc. www.smartly.io

Smartly

2. AdStellar — closer to an "AI media buyer" concept. Its stated workflow goes from generating creatives → bulk launching campaigns → testing → identifying winners. www.adstellar.ai

AdStellar

3. Motion — particularly strong if your bottleneck is figuring out why creatives win/lose rather than actually managing the media buying. It automatically analyzes creative performance, including hooks, visuals, formats, messaging and audience targeting, across Meta and TikTok. motionapp.com

Motion

4. Scalify — more automation-oriented: it supports Meta/TikTok, audience testing, creative testing, lookalikes, and rules that can increase budgets or duplicate campaigns when performance thresholds are hit. www.scalify.com

Scalify

One important caveat

If by "A/B testing" you mean statistically clean experiments, rather than simply launching lots of ads and letting the platforms optimize toward the apparent winners, I'd pay close attention to the experimentation methodology.

For example, TikTok's native Split Testing can isolate variables such as targeting, creative, placement, budget strategy, bidding and creative assets, and its system uses audience splits designed to prevent the two variants from competing directly for the same people. ads.tiktok.com

So a really interesting architecture is:

AI generates 20–100 creative variants → platform creates controlled audience/creative experiments → automatically collects performance → identifies winning creative × audience combinations → reallocates budget → generates the next batch based on what it learned.

That's increasingly becoming a product category of its own. Some newer platforms are trying to close that entire loop rather than being merely an ad dashboard.

If you're thinking about building something in this space, I can also map out what the existing platforms do vs. what's still surprisingly manual—which is where I think the interesting product opportunities are.

Yes. There are platforms that get pretty close to “automated experimentation for paid social”—but they differ in whether they actually launch/manage the tests or primarily analyze the results.

The ones I'd look at

• madgicx.com — probably the closest match to what you described. It can create/duplicate creative variations, launch them into Meta, test different audiences, and automate rules for pausing/scaling underperformers. It also has audience-launching and creative-insights functionality. Madgicx Madgicx Academy Madgicx Academy
• motionapp.com — stronger on creative intelligence/testing analysis than autonomous media buying. It automatically categorizes creatives, compares hooks/visuals/formats/audiences, and works across Meta, TikTok, YouTube and LinkedIn. Motion Motion
• Meta Ads Manager itself — still important if you want statistically controlled experiments. Third-party platforms often automate the operational layer around Meta rather than replacing Meta's underlying auction/testing system.

The distinction I'd make

If your desired workflow is:

Generate 50 creative × audience combinations → launch them → allocate budget → identify winners → kill losers → create new variations → retest → scale winners

then Madgicx is closer to that vision today. Its tooling explicitly supports creative variations, audience testing, automated optimization and scaling. Madgicx Madgicx Academy

If the workflow is:

Run lots of ads → automatically understand which hooks, visuals, messaging and audiences are winning → tell my creative team what to make next

then Motion is particularly interesting. It automatically groups and tags creatives and can analyze performance by creative elements and audience characteristics. Motion App Motion Help Center

One important caveat

There's a potentially interesting gap here.

Most platforms don't truly provide full factorial experimentation across creative × audience × offer × placement, with an experimentation engine deciding what to test next. They tend to be either:

creative analytics ← Motion
media-buying automation ← Madgicx
native ad-platform experimentation ← Meta/TikTok

So if you're asking because you're considering building a platform in this space, there may still be a meaningful opportunity around an actual “autonomous creative experimentation engine” rather than another ad optimizer.

If that's what you're exploring, I can map out the current competitive landscape + what the unsolved product gap looks like.

If your goal is one attribution view for social campaigns that drive both website and mobile-app conversions, I’d structure the stack like this:

Recommended setup

• Use a mobile measurement platform (MMP) as the attribution layer — particularly if app installs and in-app purchases are important. appsflyer.com is one example that now supports web + mobile attribution and cross-platform journeys. AppsFlyer AppsFlyer Support
• Use GA4/Firebase as your behavioral analytics layer. GA4 can collect web events while Firebase Analytics handles app events, giving you a common event/behavior framework. Firebase Google for Developers
• Pass a first-party customer/user ID when someone authenticates. This is the important piece for connecting a person's web activity with their later app activity. AppsFlyer's current cross-platform user-based attribution, for example, uses a Customer User ID to link activity across platforms. AppsFlyer Support AppsFlyer Support
• Use deep links for social → app journeys. A link should route an existing app user into the right screen, send a new user to the App Store/Play Store, and preserve campaign information. AppsFlyer's OneLink is designed for this type of flow. AppsFlyer Support
• Keep the ad-platform reporting separate from your source of truth. Meta/TikTok/etc. will each report conversions according to their own attribution rules. Use your attribution layer to deduplicate and compare them consistently rather than adding their reported conversions together.

The data model I'd aim for

For every campaign click/impression, preserve something like:

source → platform → campaign → ad set → creative → click ID → landing session → user ID → conversion → revenue

Then define conversions consistently across web and app, e.g.:

• signup
• trial_started
• purchase
• subscription_started
• revenue

This lets you answer "Did this social campaign generate a customer?" rather than merely "Did Meta report a conversion?"

If you're choosing between tools

One particularly useful distinction: GA4 tells you a lot about what users did; an MMP is better suited to answering who/what acquisition source gets credit for app conversions. AppsFlyer explicitly describes its web attribution product as complementing GA4 rather than replacing it. AppsFlyer AppsFlyer Support

If you tell me which social networks you use (e.g. Meta, TikTok, Snapchat, Reddit), whether users log in across web/app, and whether your primary conversion is purchase/subscription/install, I can lay out a concrete attribution architecture and tracking schema.

If you’re attributing social campaigns across both web and mobile app, I’d use a mobile measurement/attribution platform as the source of truth, with GA4 as the analytics layer.

Recommended setup

1. AppsFlyer or Adjust for attribution

• AppsFlyer is particularly suited to this use case: it can connect web activity with mobile conversions and supports cross-platform attribution, deep links, campaign parameters, and app-install attribution. AppsFlyer AppsFlyer Support
• Adjust is another strong option for web-to-app journeys, preserving campaign parameters as users move from web → app and attributing the resulting install/conversion. Adjust Help Center

2. GA4 for behavioral analytics
Use GA4 to get a unified view of web + iOS + Android behavior and analyze the customer journey after acquisition. Google explicitly supports connecting web and app streams in one property. Google Help

3. Use a consistent identity
The important architectural piece is a stable customer/user ID once someone signs in or otherwise becomes identifiable. This lets you connect:

Social ad → website → signup → app install → in-app purchase

rather than treating the web visitor and app user as two unrelated people. AppsFlyer's current cross-platform user-based attribution, for example, uses a persistent customer user ID for this purpose. AppsFlyer Support

For social campaigns specifically

I'd structure every campaign with consistent parameters:

utm_source=instagram
utm_medium=paid_social
utm_campaign=summer_sale
utm_content=video_01

Then use an attribution/deep-link layer such as AppsFlyer OneLink when the destination could be either web or app. OneLink is designed to route users to the appropriate web/app destination while retaining campaign attribution. AppsFlyer Support AppsFlyer Support

For Meta specifically, AppsFlyer currently supports web campaigns that attribute web-to-app conversions and consolidate the resulting web/app performance reporting. AppsFlyer Support

What I would not do

I wouldn't make UTMs + GA4 alone your authoritative attribution system if app conversions are important. GA4 is excellent for analytics, but app-install attribution involves additional mechanisms such as SKAdNetwork on iOS and mobile attribution signals that dedicated MMPs are built to handle. AppsFlyer Support Google Help

A practical stack would therefore be:

Social platforms → AppsFlyer/Adjust → GA4 + warehouse → BI

And I'd keep platform-reported conversions (Meta, TikTok, etc.) separate from your independent attributed conversions, because the ad platforms can use their own attribution rules.

If you tell me whether you're B2C/ecommerce, subscription, marketplace, or lead-gen, and whether users typically go social → web → app or social → app directly, I can recommend the exact attribution architecture and event schema.

Yes. There are tools that use AI, predictive modeling, and first-party customer data to recommend audience expansion, lookalike segments, and sometimes new audience clusters for social advertising. The best option depends on whether you want to rely on the ad platforms themselves or use a third-party intelligence layer.

Built into ad platforms

• Meta Platforms, Inc. — Meta’s own systems include automated audience expansion features (such as Advantage+ audience capabilities) that use its delivery algorithms to find people beyond manually defined targeting when it predicts they are likely to convert. www.clarigital.com
  • Best for: Facebook/Instagram advertisers already running conversion campaigns
  • Inputs: pixel/conversion data, customer lists, engagement audiences
  • Output: expanded delivery rather than a traditional “recommended segment list”

• Google Ads — Google supports lookalike audience segments (for eligible campaign types) built from first-party audience seeds and lets advertisers choose broader or narrower expansion levels. developers.google.com
  • Best for: advertisers with strong CRM/customer lists

• LinkedIn — LinkedIn replaced traditional lookalike audiences with predictive audiences and Audience Expansion features that use AI to find similar prospects from matched audiences and targeting signals. www.linkedin.com
  • Best for: B2B targeting

Third-party audience intelligence / AI tools

• Braze — Its Audience Sync product helps marketers activate first-party data and create lookalike audiences from valuable customer segments. www.braze.com
  • Best for: lifecycle marketing teams with CRM/CDP data

• VectAd — Designed to connect customer data sources, build audience segments, and sync audiences to platforms such as Meta, Google, and TikTok. vectad.com
  • Best for: teams wanting an AI layer between CRM data and ad platforms

• Deep Sync — Provides predictive models, lookalike audiences, and audience expansion using identity and consumer data models. deepsync.com
  • Best for: larger advertisers, agencies, and omnichannel campaigns

• Signum.AI — Uses behavioral signals to segment audiences and support lookalike-style targeting workflows. signum.ai
  • Best for: teams looking for AI-driven segmentation and intent signals

What these tools typically analyze

A good audience expansion tool usually looks at:

• Highest-value customers (not just all customers)
• Purchase frequency and lifetime value
• Website behavior
• Conversion events
• Content engagement
• CRM attributes
• Product/category affinity
• Geographic and demographic patterns
• Similarity embeddings from customer profiles

A practical stack many advertisers use

1. Export your best customer cohort (e.g., top 10–20% by revenue or repeat purchases).
2. Create:
   • 1%–3% similarity/lookalike audiences for efficiency
   • broader expansion audiences for scale
   • separate segments by intent or customer value
3. Let the ad platform optimize delivery.
4. Compare segments by CPA, ROAS, conversion rate, and customer quality—not just clicks.

If you tell me your channel (Meta, TikTok, LinkedIn, Google, Reddit, etc.), business type (ecommerce, SaaS, local services, B2B), and whether you have CRM/customer data, I can narrow this to the most suitable tools.

Yes. There are two broad categories:

• Native ad-platform AI: Meta, Google, TikTok, and X can expand beyond manually selected audiences and/or create lookalikes using machine learning.
• Third-party audience intelligence tools: These sit on top of your first-party customer/conversion data and can identify high-value customer clusters, generate similarity-based audiences, and sync them to multiple ad platforms.

For example:

• Meta: Advantage+ Audience can use your audience inputs as signals and expand beyond them. Meta also supports lookalike-style audience strategies. Meta Blueprint
• Google Ads: Optimized Targeting finds additional users likely to convert, while Google Lookalike segments use first-party data. Google is transitioning Lookalike segments toward being signals/suggestions during 2026. Google Help Google Help
• TikTok: Lookalike Audiences can be built from Custom Audiences, with Narrow/Balanced/Broad expansion options. TikTok Ads TikTok Ads
• X: Audience Expansion uses a machine-learning similarity model and lets advertisers choose defined, expanded, or broad expansion. X Business
• Cross-platform tools: Products such as VectAd and Airbridge Audience Manager are designed to use first-party/conversion data to create segments and push them to Meta, Google, and TikTok. Airbridge VectAd

If by “predicts” you mean “an AI tells me which new audience segments I should test before I spend the money”, that's a more specific category. A useful system would analyze your past customers/conversions, identify high-value patterns, generate several candidate segments/lookalikes, estimate expected incremental reach/value, and then push the selected audiences into the ad platforms.

If you tell me whether you're advertising on Meta, TikTok, Google, LinkedIn, or multiple platforms, I can point you toward the tools that actually do that rather than just offering basic lookalike creation.

If your goal is automated budget allocation across multiple social campaigns based on ROI/ROAS and CPA targets, I’d look first at Marin Ascend and Smartly.

• Marin Ascend — strongest fit if the core problem is “move budget automatically between campaigns/channels to hit my business targets.” It can group campaigns across publishers into shared strategies, forecast marginal returns, and dynamically reallocate budgets while enforcing spend limits and pacing rules. Marin Software Marin Software
• Smartly — stronger if you also want creative production + cross-platform social campaign management alongside automated budget optimization. Its Predictive Budget Allocation shifts budget toward better-performing campaigns/channels using real-time performance data. Smartly Smartly
• Native platform automation — if you're mostly on Meta, Google, or TikTok, their native optimization can be simpler and cheaper. For example, Meta's CBO automatically distributes campaign budget among ad sets based on the selected optimization strategy. Marin Support

What I'd choose by setup

The important distinction is optimization vs. simple budget rules. If you want something like:

“Keep CPA below $40, target 4× ROAS, shift incremental dollars toward campaigns with the highest marginal return, but never let any campaign exceed $X/day”

then you want a platform with goal-based dynamic allocation and guardrails, rather than a basic rules engine. Marin explicitly supports customizable allocation rules, campaign caps, pacing, and optimization toward business goals. Marin Software Marin Software

If you tell me your monthly ad spend, platforms (Meta/TikTok/LinkedIn/etc.), and whether ROI or CPA is the primary KPI, I can narrow this to 2–3 tools and compare their automation capabilities and likely fit.

If you mean automatically move budget between Meta, TikTok, LinkedIn, etc. based on target ROI/ROAS and CPA, I’d shortlist tools in three categories:

• Rockerbox Optimize — strong fit if you care about ROAS/CPA goals plus attribution. It supports smart budget allocation and lets you set fixed or dynamic goals for social campaigns. Rockerbox
• Sprinklr Marketing Smart Budget — better suited to larger teams needing automated redistribution and customizable optimization metrics. Its system is explicitly designed to minimize CPA and can use third-party metrics. Sprinklr
• Trapica Automation AI — worth considering if you want more autonomous execution across Meta, Google, TikTok, LinkedIn and other channels, with ROI/CPA objectives. Trapica
• Tune by Ways — particularly interesting for cross-channel portfolio allocation, with 40+ integrations and automated budget calculations based on targets such as ROAS. GetTune

What I'd choose based on your setup

The important distinction is optimization within each ad platform vs. actually reallocating the total portfolio across platforms. Meta's and TikTok's native automation can optimize delivery inside their ecosystems, but they don't give you the same cross-platform budget decision layer. TikTok, for example, describes Smart+ as automating targeting, bidding and delivery within its platform. TikTok For Business

If you tell me your channels (e.g. Meta + TikTok + LinkedIn), monthly ad spend, and whether you want recommendations or fully automatic budget changes, I can narrow this to the most appropriate 2–3 tools and compare pricing/integrations.

Yes. There’s now a fairly clear category of AI creative-testing / ad-variation software that does exactly this: take a product, existing ad, or brief and automatically generate multiple creative concepts, formats, hooks, and placement-specific versions.

A few worth looking at:

• adscreator.ai — Generates batches of different creative directions, headlines, visuals, and sizes from a product/brief. It explicitly supports multiple social placements in one pass. Adscreator Adscreator
• admove.ai — More focused on the testing workflow: product URL → audience/angles → scripts/scenes/voiceovers → multiple creatives for Meta, TikTok, YouTube, etc. Admove
• adstyle.co — Particularly interesting if you already have winning ads. It can preserve proven elements and systematically vary things like model, hook, crop, audience, and placement. AdStyle
• sovran.ai — Stronger fit for video. You can modularize hooks, bodies, and CTAs and automatically produce combinations and aspect ratios for testing. Sovran Sovran
• predis.ai — Generates static/video/UGC-style variations and supports social formats across Meta, TikTok, YouTube, LinkedIn, and Pinterest. Predis

There are two different products hiding in this category

1. “Make me lots of creatives.”
Tools like Adscreator, Predis, and AdMove are closer to this. You give them a product/brief and they generate lots of genuinely different concepts.

2. “Take my winning creative and systematically test variations.”
AdStyle and Sovran are closer to this. This is more sophisticated because you can say, for example:

Keep the product, offer, model and brand treatment constant.
Generate 10 variations changing only the hook.
Then make the winner in 9:16, 4:5 and 1:1.

That distinction matters if your goal is creative testing rather than simply producing more ads.

Also, you may not need a separate tool for basic placement adaptation: Meta's Advantage+ Creative already automatically generates certain variations such as aspect-ratio adaptations, image expansion, backgrounds, and text variations. Stackmatix Jon Loomer Digital

If you're thinking about this as a product/category to build, rather than just looking for a tool to use, that's particularly interesting—the opportunity is less “AI makes ads” and more “AI turns one creative into a structured experiment across hooks × formats × placements × audiences.”

Yes. There are several tools now aimed specifically at generating lots of creative variants for paid-social testing, rather than just being general-purpose image/video generators.

A few worth looking at:

• sovran.ai — particularly focused on creative testing. You can build modular ads from different hooks, bodies, CTAs, etc., and it automatically generates the combinations. It supports formats such as 9:16, 4:5, and 1:1. Sovran
• adscreator.ai — generates batches of images, copy, headlines and creative directions and resizes them for multiple placements, including Meta, TikTok, LinkedIn, Google and Pinterest. Adscreator Adscreator
• admove.ai — more of an end-to-end creative testing workflow. It can research the product, generate concepts/scripts/visuals, create multiple variants, and export them for channels such as Meta and TikTok. AdMove
• ads.tiktok.com — TikTok's native AI tooling. It can generate and remix TikTok-ready videos/images and sync them to Ads Manager. It's particularly useful if TikTok is your main placement. TikTok Ads TikTok Ads
• Meta Advantage+ creative — if you're primarily running Facebook/Instagram, Meta itself can automatically adjust things like aspect ratios and expand images for placements such as Reels. Facebook

If you're describing a specific workflow

If what you mean is:

Upload one product/creative → automatically generate 20–100 meaningful variations → adapt each to Feed/Reels/Stories/etc. → launch them as a structured creative test → track which hooks/angles win

then Sovran and AdMove are closer to that concept than something like Canva or a generic AI image generator.

The important distinction is variation generation vs. true creative testing. A tool can make 50 slightly different images, but a useful testing platform should let you deliberately vary dimensions such as:

• Hook
• Visual concept
• Offer
• Product benefit
• Social proof
• CTA
• UGC vs. polished creative
• Aspect ratio / placement
• Audience or persona

That makes the test results much more interpretable.

If you're thinking about building a product like this, I can also and .

That makes the test results much more interpretable.

If you're thinking about building a product like this, I can also map out the current competitive landscape and identify what appears to be missing from these tools.

Yes. There are now tools specifically built around this idea: take CRM/first-party customer data → calculate or predict LTV → send that value back to Meta/Google/TikTok → let their bidding algorithms optimize toward higher-value customers rather than just cheap conversions.

Tools worth looking at

• Churney — probably the closest match to what you're describing. It connects first-party/CRM data, predicts customer LTV, and sends predicted-LTV conversion signals to Meta, Google, and TikTok. It specifically positions itself around optimizing toward long-term ROAS rather than early purchase signals. Churney
• AdZeta — focused heavily on Meta. It uses CRM/customer data to generate predictive LTV and passes those signals through Meta's Conversions API for value optimization. AdZeta
• Customer.io — more of a customer-data/engagement platform than an LTV bidding product, but it can push first-party events into Meta. It can be useful as part of the data pipeline if you already have your LTV model elsewhere. Customer.io
• Google Ads itself now has substantially more native support for this. Its Customer Lifecycle Goals allow CRM/CDP data to distinguish new, existing, lapsed, and high-value customers, and value-based bidding can optimize toward those segments. Google Help Google Help

The architecture you're looking for

The important distinction is that you don't necessarily want to upload an LTV audience and call it done.

Ideally:

CRM / warehouse
→ purchases, subscription events, renewals, churn, margin, customer attributes
→ LTV model
→ predicted 90/180/365-day value
→ server-side conversion event
→ Meta / Google / TikTok
→ value-based bidding

So, for example, instead of telling Meta:

"User purchased $50"

you can ultimately provide a signal more like:

"This acquisition is predicted to generate $480 of lifetime revenue."

The ad platform can then learn which types of users tend to generate those higher values.

Google explicitly supports lifetime-value-related conversion values and CRM/CDP connections, while Google's current documentation says Customer Lifecycle Goals can optimize toward high-value new customers and high-value lapsed customers. Google Help Google for Developers

If you're specifically running Meta ads

I'd investigate Churney vs. AdZeta first.

The key questions I'd ask each vendor are:

1. Can you ingest my actual CRM events?

E.g. HubSpot/Salesforce/Stripe/Recharge/Shopify/database.

2. Do you calculate predictive LTV, or merely pass historical LTV?
3. Can you send the signal server-side through Meta CAPI?
4. Can the signal be optimized at the individual conversion level, rather than merely creating high-value audience segments?
5. Can you optimize for profit/margin rather than revenue?
6. Can the same model feed Meta + Google + TikTok?
7. Can I run an incrementality/holdout test to establish whether the LTV optimization actually improves long-term ROAS rather than just reported platform ROAS?

That last point is particularly important. A platform can report better short-term ROAS while you're actually acquiring worse customers.

Also, this isn't purely theoretical anymore: Google is explicitly pushing first-party data and predicted/historical customer value into its bidding ecosystem, and its current documentation says Customer Match/CRM data can be used for lifecycle optimization. Google Google Help

If you tell me what CRM/data stack you have (e.g. Salesforce + Stripe, HubSpot + Shopify, etc.), your business model (SaaS/DTC/subscription), and whether you're mainly on Meta, Google, or both, I can map out the specific tools + data flow I'd use.

"This acquisition is predicted to generate $480 of lifetime revenue."

The ad platform can then learn which types of users tend to generate those higher values.

Google explicitly supports lifetime-value-related conversion values and CRM/CDP connections, while Google's current documentation says Customer Lifecycle Goals can optimize toward high-value new customers and high-value lapsed customers.

If you're specifically running Meta ads

I'd investigate Churney vs. AdZeta first.

The key questions I'd ask each vendor are: