Frequently Asked Questions

BigAtom is a Product Performance Management (PPM) platform built specifically for e-commerce brands running catalog ads on Meta and Google. Rather than just managing feed data, BigAtom connects feed quality to ad performance outcomes — with built-in analytics, automation, and creative tools.

Core strengths:

• SKU-level performance analytics across Meta and Google
• Creative automation (dynamic overlays) for catalog ads
• Smart product segmentation (bestsellers, hidden gems, under-performers)
• Stop-loss automation and broken inventory management
• Built specifically for performance-focused e-commerce brands

Primary focus: Catalog ad performance optimization

Traditional ad management tools answer the question: “How are my campaigns performing?”

A Product Performance Management platform answers: “Which products in my catalog are driving growth, which are wasting budget, and what should I do about each one?”

The unit of analysis shifts from the campaign to the product (SKU). Everything — analytics, automation, creative, budget allocation — is organized around individual products rather than campaign structures.

A PPM platform provides:

• SKU-level performance analytics — aggregated from all channels at the product level
• Intelligent product segmentation — auto-classifying products into performance tiers
• Feed management and optimization — maintaining and enriching product feeds
• Creative automation — dynamic overlays applied at catalog scale
• Budget automation — scaling winners, suppressing under-performers
• Inventory health management — removing broken inventory from active campaigns

A PPM platform becomes essential when:

• Your catalog exceeds 200+ SKUs — Manual SKU-level management becomes impractical
• You are spending Rs.5+ lakh/month on catalog ads — The efficiency gains from automation justify the investment many times over
• You are running on multiple channels — Meta + Google + others requires a unified product view
• Your inventory changes frequently — Fast-moving catalogs need real-time monitoring
• You are trying to scale profitably — Indiscriminate scaling without product-level intelligence leads to deteriorating ROAS

Feedonomics is a feed management platform focused on feed optimization, transformation, and syndication across 100+ channels. BigAtom is a Product Performance Management platform focused on catalog ad performance optimization.

Choose BigAtom if your primary channels are Meta and Google and you want performance analytics plus automation, not just feed management.

Choose Feedonomics if you need to distribute your feed across 20+ channels globally or are a large enterprise with complex feed transformation requirements.

DataFeedWatch asks: “Is your feed data correct and properly formatted for each channel?”

BigAtom asks: “Are you getting the best possible performance from your catalog ads, and what should you do about each individual product?”

Choose BigAtom if Meta is a primary ad channel and you want SKU-level performance visibility, automation, and creative tools.

Choose DataFeedWatch if you primarily need Google Shopping feed optimization or feed distribution across a large number of niche channels.

A product feed (also called a data feed) is a structured file — usually in XML, CSV, or JSON format — that contains all the information about your products. It includes:

• Product title
• Description
• Price and sale price
• Images
• Category and product type
• SKU or item ID
• Availability and stock levels
• URLs

Ad platforms like Google Merchant Center and Meta Commerce Manager use this feed to dynamically generate ads for each product in your catalog. The quality of your feed directly determines the quality, relevance, and performance of those ads.

Product feed management is the process of creating, optimizing, distributing, and maintaining your product feeds across advertising and shopping channels. It involves:

• Feed creation: Pulling product data from your e-commerce platform
• Feed optimization: Enriching and cleaning data to meet channel requirements
• Feed distribution: Sending the right data to the right channels
• Feed monitoring: Catching errors, disapprovals, and data quality issues
• Performance analysis: Understanding which products are performing and why

Done well, feed management is not a one-time task. Brands that invest in it typically see a 15-30% reduction in wasted ad spend and 10-25% improvement in CTR.

Many brands manage feeds manually — exporting spreadsheets, making changes, and re-uploading. This works at small scale but breaks down fast as catalogs grow.

Manual management problems:

• Time-consuming and error-prone
• Cannot react to real-time inventory changes
• No performance-based automation
• No creative customization at scale

Automated feed management handles feed optimization, segmentation, and suppression automatically — using actual performance data to make decisions at the SKU level. Budget naturally flows to products that convert, without manual intervention.

Facebook catalog ads dynamically pull products from your Meta Commerce Manager catalog and show them to users based on their browsing behavior, purchase history, or interests. Every user can see a different product, personalized to what they are most likely to buy.

They are used for:

• Retargeting — showing products to people who viewed or added them to cart
• Prospecting — finding new customers similar to your existing buyers (Advantage+ Catalog Ads)
• Cross-selling — showing complementary products to past purchasers

The most common reason catalog ads underperform is that brands connect their product catalog and let Meta’s algorithm do all the work — without taking control of what gets promoted and how it looks.

Advantage+ Catalog Ads use Meta’s AI to determine who to show your products to, without requiring you to specify a retargeting audience or custom audience pool. Meta’s system analyzes your catalog, looks at user behavior signals, matches products to users most likely to purchase, and automatically optimizes over time.

Common mistakes to avoid:

• Evaluating too early — Do not pause campaigns in the first 7 days
• Running with too small a catalog — Under 50 products limits what the algorithm can learn
• Using low-quality images — Kills performance regardless of targeting quality
• Not excluding out-of-stock products — Wastes impressions and frustrates users
• Combining prospecting and retargeting — Makes it impossible to evaluate either effectively

Dynamic Product Ads (DPA) are catalog-driven, automatically generated ads. They pull product data from your feed and personalize the ad for each viewer. Best for brands with large catalogs (50+ SKUs), strong retargeting audiences, and high-frequency shoppers.

Static Ads are manually created with consistent creative. They work best for brand awareness, new product launches, and premium or luxury positioning where brand story matters as much as the product.

The winning strategy is to combine both: DPA for retargeting and broad catalog coverage, static ads for brand building and hero product moments.

Many high-performing brands run both: Standard Shopping for their top 20-30% of products where maximum control is needed, and Performance Max for the broader catalog.

1. Running the entire catalog without segmentation — Segment by performance tier, margin, and inventory health.
2. Ignoring product feed quality — Generic titles, missing attributes, and low-res images hurt ROAS by 20-30%.
3. Not excluding out-of-stock and broken inventory — Set up automatic exclusion rules.
4. Using raw product images with no creative treatment — Add overlays; even simple ones improve CTR 15-25%.
5. Judging performance too early — Minimum 14-30 days and 500+ impressions per product before deciding.
6. Not monitoring at the SKU level — Pull product-level performance breakdowns weekly.
7. Missing or incorrect GTINs on branded products — Source and include correct UPC, EAN, or ISBN.
8. Confusing retargeting and prospecting results — Separate into distinct campaigns with separate budgets.
9. Not re-evaluating paused products — Review paused products every 30-60 days.
10. Optimizing for ROAS without considering margin — Incorporate margin data into your optimization framework.

Avoiding these 10 mistakes can realistically improve catalog ad ROAS by 30-50% without increasing your budget.

Stop loss automation is a system that monitors individual product performance in real time and automatically pauses, suppresses, or deprioritizes products that fall below your defined performance thresholds.

Common stop loss rules include:

• Spend-based: Pause any product where spend exceeds 3x the product’s average order value without a purchase
• ROAS-based: Pause any product with a 7-day ROAS below your minimum threshold
• Conversion-rate: Pause products with 100+ clicks but zero add-to-carts
• Impression-based: Flag products with very low impressions despite being in active campaigns

Most brands see measurable improvement within the first 30 days of implementing stop loss automation.

That is why good stop loss systems include grace periods before pausing, re-evaluation windows so paused products get a second look, and exceptions for new products. Start conservative: use generous thresholds (higher spend limits before pausing). Products paused for underperformance should be reviewed every 30-60 days to see if conditions have changed — new pricing, new inventory, or seasonal relevance.

Algorithms work better with concentrated, consistent conversion signals. Spreading budget thin across 500 products — including many that never convert — actually gives the algorithm worse data, not better. By pausing non-converting products, you concentrate spend on products where the algorithm can gather meaningful conversion data, leading to faster optimization and better performance overall.

Creative automation uses templates and data rules to automatically generate customized visuals across your entire product catalog. You design a template once, connect it to your product data, and apply it across your entire catalog automatically.

Types of overlays that drive performance:

• Discount and offer badges — The most impactful single overlay element
• Price display overlays — Communicate value directly in the feed
• New arrival tags — Capture interest in fresh inventory
• Urgency and scarcity indicators — Drive action with limited-stock messaging
• Brand and category labels — Reinforce brand identity at scale
• Rating and review overlays — Add social proof to product images

Even simple overlays typically improve CTR by 15-25%.

Broken inventory refers to products that have incomplete or severely limited variant availability — a shirt only in XXL, shoes only in Size 5 and Size 11, a dress in one color when it originally came in five.

The product is technically in stock so it does not get filtered by standard out-of-stock rules. But in practice it is nearly unconvertible. When a customer clicks through and discovers only XXL is available (and they are a Medium), you have paid for a click with zero chance of converting. BigAtom automatically identifies and excludes these products from active campaigns.

A hidden gem is a product that has high ROAS or high conversion efficiency but receives relatively low impressions or ad spend — profitable when shown, but not being shown enough.

How to identify hidden gems:

1. Pull SKU-level performance data (impressions, clicks, ROAS per product)
2. Calculate a Performance Efficiency Score (ROAS x CTR divided by impression share)
3. Filter for high efficiency score plus low impression share
4. Cross-reference with margin data to prioritize the most profitable opportunities

ROAS measures the revenue generated for every rupee spent on advertising:

Example: If your catalog campaign spent Rs.1,00,000 and generated Rs.4,00,000 in revenue, your ROAS is 4x. ROAS is a revenue metric, not a profitability metric. For a more complete picture:

20% margin means Break-Even ROAS = 5x. 50% margin means Break-Even ROAS = 2x.

These are starting benchmarks, not universal targets. Know your margins and set your targets accordingly.

SKU-level analytics measures advertising performance at the individual product level rather than just at the campaign or ad set level. Instead of knowing that your “Men’s Apparel” campaign returned 3.2x ROAS, you know that SKU #1042 returned 7.8x and should get more budget, while SKU #2291 spent Rs.12,000 with zero conversions and should be paused.

A campaign can look fine at 3x ROAS while 100 products burn through budget with zero conversions, subsidized by 50 products that are genuinely performing. SKU-level data reveals exactly where that is happening.

The ROAS number your ad platform reports depends heavily on attribution settings:

• 7-day click + 1-day view attribution (Meta default): Counts conversions from views, which inflates ROAS
• 7-day click only: More conservative; better for evaluating catalog ad efficiency
• Data-driven attribution (Google): Distributes credit across touchpoints; more accurate for multi-channel brands

Always verify which attribution model you are using before drawing conclusions from ROAS data. Retargeting campaigns typically return 5-10x ROAS; prospecting might return 2-3x. Keeping them separate lets you optimize each appropriately.

1. Exclude non-converting products — The fastest ROAS improvement lever.
2. Concentrate budget on high-ROAS products — Even a 20% budget shift to proven winners typically improves overall ROAS significantly.
3. Add creative overlays — Higher CTR from discount badges and price labels improves Quality Scores and lowers effective CPC.
4. Separate retargeting and prospecting — Lets you optimize each appropriately.
5. Fix broken inventory — Clicks on broken inventory products have near-zero conversion probability.
6. Improve landing pages — If click-through rates are good but conversion rates are low, focus on page load speed, product imagery, and size availability.
7. Adjust bids by margin — Higher bids for high-margin products, lower for low-margin.
8. Improve feed quality — Better product titles, accurate pricing, and complete attributes improve match quality on Google and relevance on Meta.

Meta’s Offline Conversions API (OCAPI) — now unified under the broader Meta Conversions API (CAPI) — is a server-to-server data integration that allows advertisers to send offline conversion events (in-store purchases, phone orders, CRM leads, walk-in sales) directly to Meta’s ad infrastructure. This closes the attribution loop between digital ad spend on Facebook and Instagram and real-world business outcomes.

Brands that feed offline signals into Meta’s AI engine see dramatically better campaign optimization and ROAS — typically 15-30% improvement in attributed conversions.

As of May 2025, the standalone Offline Conversions API is deprecated. Offline events are now sent via the unified Meta Conversions API using action_source: “physical_store”. The key differences between legacy and current:

Meta’s Event Match Quality (EMQ) score determines how effectively your offline events are matched to real Meta users. Higher EMQ means more accurate attribution and better algorithmic optimization.

Target EMQ scores by event type:

• Purchase events: 8.8 to 9.3
• Lead events: 7.5 to 8.5
• PageView events: 6.5 to 7.5

The most common reason for low EMQ is missing or incorrectly normalized PII. Sending additional identifiers (name, city, ZIP) alongside hashed email and phone significantly improves match rates. Active monitoring and data enrichment recommendations are required — simply uploading data without EMQ optimization typically delivers only a fraction of the potential improvement.

Offline events should ideally be sent within 24-48 hours of the conversion. The legacy system allowed a 62-day upload window, but real-time CAPI submission is strongly recommended for optimal matching.

Uploading stale data (weekly batch uploads instead of daily automated uploads) introduces conversion lag that confuses Smart Bidding and reduces EMQ scores. Near-real-time or at minimum daily automated uploads are the standard for best results.

At minimum, a hashed email or phone number is required. Adding name, city, and ZIP significantly improves match quality and EMQ scores.

All customer identifiers must be SHA-256 hashed after normalization before sending:

• Email: Convert to lowercase, remove leading/trailing whitespace, then hash
• Phone: Remove all non-numeric characters, include country code (e.g., 919810001234 for India), then hash
• Name, city, ZIP: Normalize and hash — improves EMQ significantly

Sending raw (unhashed) PII violates both Meta’s platform policies and privacy regulations (GDPR, PDPB, CCPA).

Yes. Integration partners can connect your Shopify, WooCommerce, Magento, or custom e-commerce platform with CAPI to automatically route both online and offline events to Meta. This includes:

• Online events: Purchase, AddToCart, ViewContent from your website server
• Offline events: In-store purchases via POS integration, COD delivery confirmations, WhatsApp-assisted sales
• CRM events: Lead status changes, deal closed events routed via action_source: “system_generated”

Native platform integrations handle SHA-256 hashing, normalization, and deduplication automatically, removing the need for custom development.

Retail and D2C Brands: Track in-store purchases linked to Meta ad clicks. Measure the true omnichannel ROAS by connecting your POS data to Meta.

Lead Generation Businesses: If your sales cycle spans days or weeks (real estate, insurance, EdTech), push CRM conversion events — “Deal Closed”, “Policy Sold” — back to Meta when they occur.

Quick Service Restaurants (QSR): Connect food delivery or in-restaurant order data with Meta campaigns running awareness and offers.

Automotive and Financial Services: Link vehicle purchase or loan disbursement events — happening weeks after the initial ad click — to the original campaign for true attribution.

Indian D2C Specifics: COD order delivery confirmation events, WhatsApp commerce attribution, and loyalty card scan events are uniquely valuable signals for Indian brands that standard pixel tracking completely misses.

The legacy Offline Conversions API was discontinued on May 14, 2025. Migration to unified CAPI involves these key steps:

1. Locate your Dataset ID in Meta Events Manager — this replaces the old Offline Event Set ID
2. Add the action_source parameter to all offline events: use “physical_store” for in-store purchases, “system_generated” for CRM events, “phone_call” for phone orders, “chat” for WhatsApp-assisted sales
3. Implement event deduplication using the event_id parameter to prevent double-counting against Pixel events
4. Add data_processing_options — this is now mandatory for privacy compliance
5. Switch to the standard CAPI endpoint — no more separate Offline Event Set endpoint
6. Validate EMQ scores in Events Manager — target 8.0+ for Purchase events
7. Update campaign attribution settings and bid strategies to leverage the newly unified offline signal

Google Store Sales Integration (SSI) — also referred to as Google Store Sales Improvements (GSSI) — is Google’s flagship online-to-offline ad measurement solution. It quantifies the direct impact of digital ad spend on in-store and offline purchases by matching first-party transaction data (from your CRM, POS, or loyalty program) with Google users who were exposed to or clicked on your ads.

SSI is powered by Google’s AI and machine learning, using a rich combination of data signals — ad interactions, store visits, Google Opinion Rewards surveys, and third-party data partnerships — to produce privacy-safe, aggregated reports on your omnichannel ROAS.

Without SSI, Google Ads only sees what happens online — optimizing toward incomplete signals and undervaluing campaigns that drive real-world purchases. Brands using SSI with Smart Bidding consistently report 20-40% improvement in omnichannel ROAS compared to online-only measurement.

Store Sales Direct was the beta name for the program. The current official program is called Store Sales Measurement / Store Sales Integration (SSI) or GSSI, combining both automated and manual upload options under one umbrella. The terms SSI, GSSI, and Store Sales Improvements are all used interchangeably to refer to the same Google Ads program.

GSSI is a Google-paid program — there is no additional charge to advertisers for participation. However, working with a certified data partner to automate data pipelines and manage ongoing uploads may have associated service costs. The investment in a certified SSI partner is generally more than offset by the ROAS improvement from accurate Smart Bidding signals.

Google uses hashed customer identifiers (email, phone) submitted with your transactions and matches them to logged-in Google users who previously clicked on your ads. Results are aggregated and extrapolated to represent an estimate across all clicks.

The end-to-end process:

1. User clicks on Google Ad
2. User visits physical store
3. User completes in-store purchase (recorded in POS / CRM)
4. Advertiser exports transaction data with hashed PII
5. Data uploaded to Google Ads API (via direct upload or certified SSI partner)
6. Google matches transaction to ad click
7. Offline conversion reported in Google Ads and unlocks Smart Bidding on real purchases

Results are never exposed at an individual user level — all reporting is aggregated and privacy-safe.

Matching and reporting typically takes 24-72 hours after data upload. Conversion lag reporting is available to track sales that occur days or weeks after the ad click.

This is why daily automated uploads are strongly recommended over weekly batch uploads — delayed data reduces the quality of Smart Bidding signals and slows down campaign optimization.

You may not qualify for the full SSI program, which currently requires a minimum of 300,000 store visits in the past 90 days, active Location Extensions, and is available for Retail and QSR verticals only (subject to Google whitelisting).

However, if you do not qualify for SSI, you can still implement Google Offline Conversion Imports (OCI) for lead-based businesses, which has no store visit requirement. OCI allows you to upload CRM events (lead converted, deal closed, application approved) and attribute them back to Google ad clicks.

A certified implementation partner can assess your SSI eligibility and manage the Google whitelisting process on your behalf.

One of the most powerful outcomes of SSI is enabling Google’s Smart Bidding algorithms to optimize toward real in-store revenue, not just online conversions:

• Target ROAS (tROAS): Bid toward a return on ad spend that accounts for both online and offline purchases
• Target CPA: Optimize cost-per-acquisition when the acquisition is an in-store transaction
• Value-Based Bidding: Allocate budget toward ad clicks most likely to result in high-value offline purchases

Before SSI, Google’s algorithm was optimizing toward a partial view of your business — online conversions only. With SSI data, the bidding algorithm sees your full omnichannel picture and can make significantly better decisions about where to spend your budget.

First-party data is information that your customers voluntarily share with you directly — email addresses from orders, phone numbers linked to loyalty cards, purchase history, browsing behavior on your website, and shipping records. This is your data. You collected it directly. Customers consented to share it with you.

Third-party data (collected by others and shared with advertisers) has become increasingly restricted — by Apple’s iOS privacy changes, Google’s cookie phase-out plans, and regulations like GDPR, India’s PDPB, and CCPA.

As third-party data erodes, first-party data becomes the most valuable asset a brand can own. Without it, your ad platforms see fragments of your actual business — iOS 14+ blocks 30-60% of iOS conversion signals, ad blockers affect 25-40% of users, and all offline purchases are completely invisible. The result is structural underperformance: campaigns optimizing toward a small slice of your actual customers.

It depends on your media mix. If you are running significant spend on both Meta and Google, implementing both ensures you are not leaving attribution gaps on either platform. Most omnichannel brands benefit from both.

Running both in a unified first-party data strategy is transformational: no attribution conflicts, consistent audience signals across platforms, cross-platform budget optimization, and one view of omnichannel performance.

Basic CAPI and SSI implementation is typically completed within 2-3 weeks. Full campaign restructuring and optimization takes 4-6 weeks to show measurable ROAS impact.

Typical implementation timeline:

• Week 1-2: Audit existing tracking setup, map offline data sources (POS, CRM, loyalty), design unified data architecture
• Week 3-4: Build automated data pipelines, configure SHA-256 hashing and deduplication, validate event receipt and match quality
• Week 5-6: Restructure campaigns for offline-signal-enhanced bidding, enable Smart Bidding with SSI data, launch value-based bidding experiments
• Week 7+: Monitor EMQ and match rates weekly, expand offline signals to additional event types, scale budgets toward highest offline-ROAS campaigns

At minimum you need: customer email or phone at the time of offline transaction, transaction value, and transaction timestamp. Richer data enables better segmentation and optimization.

For Meta CAPI offline events:

• Hashed email and/or phone (mandatory)
• Hashed first/last name, city, state, ZIP, country (improves EMQ significantly)
• Transaction value and currency
• Event timestamp (Unix format)
• action_source value (physical_store, system_generated, phone_call, chat)

For Google SSI:

• Hashed customer email and/or phone
• Transaction amount and date/time
• EEA user consent signals (required for EU)
• Optional: product category, store location, custom business dimensions

1. Sending unhashed PII — All customer identifiers must be SHA-256 hashed. Sending raw PII violates platform policies and privacy regulations.
2. Ignoring deduplication — If you are running both Meta Pixel and CAPI, ensure deduplication is properly configured using event_id. Otherwise conversions are double-counted, inflating reported performance.
3. Uploading stale data — For Meta CAPI, offline events should be sent within 24-48 hours. For Google SSI, daily automated uploads are the standard. Weekly batch uploads confuse Smart Bidding.
4. Treating data upload as the finish line — Getting data into the platform is step one. Restructuring campaigns to leverage offline signals — adjusting bidding strategies, updating audience segments, refreshing creative — is where the ROAS improvement actually comes from.
5. Using a data tool without campaign expertise — Connecting a CAPI tool doesn’t automatically improve your campaigns. The offline data needs to be used in bid strategies, audience segmentation, lookalike creation, and creative decisions.

Indian D2C brands have unique first-party data signals that standard pixel tracking completely misses:

• COD Delivered Revenue: A large portion of Indian D2C orders are COD. The actual conversion happens at delivery — not at order placement. Sending a “COD delivered” event back to Meta via CAPI gives the algorithm a real revenue signal instead of an intent signal.
• WhatsApp and Phone Order Attribution: High-AOV categories (jewelry, home decor) frequently complete via WhatsApp or phone. These sales are completely invisible to standard tracking. Routing these as action_source: “chat” or “phone_call” events via CAPI closes this gap.
• In-Store Purchase Attribution: Brands with physical stores can attribute Meta ad clicks to in-store purchases using POS integration — enabling true omnichannel ROAS measurement.

Indian D2C brands that implement CAPI for these three signal types typically see 20-35% more attributed conversions in Meta and significantly better campaign optimization as a result.

Datahash is a no-code CDP focused on data implementation. AdYogi is a performance marketing platform and managed services partner — a fundamentally different scope.

Choose AdYogi if you need data infrastructure AND campaign execution — a partner accountable for ROAS outcomes, not just data upload success rates.

Choose Datahash if you only need basic web CAPI setup and have a separate agency managing campaigns.

Getting offline data into the platform is step one, but it does not automatically improve ROAS. Consider this common scenario:

1. You upload offline conversion data via CAPI or SSI — done
2. Your Event Match Quality scores are healthy — done
3. But your campaigns are still structured poorly — problem
4. Your bidding strategy does not account for offline LTV — problem
5. Your catalog is not optimized for high-AOV products — problem

Result: Clean data flowing into a suboptimal campaign structure delivers only marginal improvement. The real value comes from using the offline data to restructure campaigns, enable value-based Smart Bidding anchored to omnichannel LTV, create lookalike audiences from offline buyers, and reallocate budget toward the highest offline-ROAS campaigns.

Stape: A cost-effective server-side tagging infrastructure (Google Tag Manager server container). Best for brands with strong in-house development teams who want flexible, affordable CAPI gateway infrastructure. No campaign management, no India-specific features, not a managed service.

LeadsBridge: Best for lead generation businesses (real estate, EdTech, BFSI, insurance) that want to sync CRM data to Meta and Google for offline lead event tracking. Wide CRM integration library (Salesforce, HubSpot, Zoho, 300+ others) but not specialized for e-commerce or D2C use cases.

Datahash: Fast, no-code Meta Web CAPI setup. Good for brands that only need web CAPI and have separate campaign management in place. Base plan is web CAPI only — offline CAPI and Google SSI require plan upgrades.

Based on implementations for D2C and retail brands across India and globally, brands that deploy Meta CAPI (Offline) + Google SSI with a full integrated approach typically see:

Results vary based on business category, data quality, and campaign structure. Brands that only implement the data pipeline without restructuring campaigns see a fraction of these improvements.