How to Choose a Lead-Generation Agency That Protects CAC

Written by: Aaron Rovner, Founder, Saas Hero | Last updated: August 29, 2026

Key Takeaways

• Use a seven-criteria scorecard that focuses on CAC protection and payback period instead of lead lists or channel comparisons.
• Compensation model, attribution window, conversion architecture, CRM sync, post-click ownership, strategy cadence, and a 90-day validation gate directly shape CAC and payback.
• Agencies need to own landing pages, run a proactive strategy, and rebuild conversion tracking at onboarding so accountability and data quality start on day one.
• Primary conversion events should be SQL or opportunity creation, with offline CRM imports training the bidding algorithm on revenue instead of raw form fills.
• Apply this scorecard to your current account with SaaSHero and find CAC and payback gaps before you sign a new agency contract.

What a CAC-Optimized Lead-Gen Agency Actually Does

A CAC-optimized lead-gen agency structures its compensation model, attribution architecture, conversion hierarchy, CRM integration, post-click ownership, strategy cadence, and validation mechanics to reduce the cost of acquiring a revenue-generating customer, not just a form fill. Every operational decision connects back to closed-won revenue instead of lead volume.

The table below shows how to evaluate each of these seven criteria in practice, with specific metrics that prove whether an agency meets the standard.

1. Compensation Model That Removes Spend Conflicts

The first criterion to evaluate is compensation model because the agency’s fee structure determines whether its recommendations support your CAC goals or its own revenue targets. The agency’s fee structure must remove any financial incentive to grow spend or expand channel count for its own sake.

A percentage-of-spend retainer creates a structural conflict at the center of the relationship. Every recommendation to scale carries an undisclosed interest, and every recommendation to cut spend reduces the agency’s revenue. A healthy sourced pipeline multiplier for lead-gen agency spend is 8× or higher, a number that is impossible to defend honestly when the agency earns more by inflating the denominator.

A per-channel fee creates the same problem in a different form. If each additional channel raises the invoice, the agency has a financial reason to keep the channel mix exactly as it is, which causes budget to calcify where it was first placed long after the opportunity has moved. A flat retainer indexed to total monthly ad spend removes both conflicts by decoupling the agency’s revenue from any single channel decision, so every recommendation can be argued on evidence instead of invoice impact.

SaaSHero charges a flat monthly retainer set against total ad spend under management. Adding a channel, closing one, or shifting budget between them does not change the fee, so channel-mix recommendations arrive as strategic proposals rather than contract negotiations.

Questions to ask any agency:

• Does your fee change if we add or remove a channel?
• Does your fee change if we increase or decrease total spend?
• Have you ever recommended a client reduce spend, and what happened to your invoice?

Red flags:

• Percentage-of-spend pricing with no spend floor or ceiling
• Per-channel line items that require a contract amendment to test something new
• Inability to name a client they recommended reduce spend

Metric proving this criterion is met: the agency can recommend pausing a channel or cutting budget without any change to its own revenue.

See how SaaSHero’s spend-indexed retainer removes fee conflicts from every channel-mix decision.

2. Attribution Window Aligned With Your Sales Cycle

The next criterion is attribution window, because even a clean fee model fails if the agency measures performance on a fraction of the revenue your ads create. The attribution window must be long enough to capture the full sales cycle, or the channels that create demand appear worthless and get defunded.

The median B2B SaaS sales cycle is 84 days from first touch to first charge, while the median attribution window teams use is 27 days, which causes 40% of attributable revenue to be missed or miscredited. The standard rule is to set the lookback window to at least 1.5× the average sales cycle length, so a 60-day cycle warrants a 90-day window and a 120-day cycle warrants 180 days.

Extending the attribution window from 30 days to 90 days raises attributed revenue from 58% to 87%, and extending to 180 days raises it to 94%. Using Meta’s default 7-day click plus 1-day view windows in B2B systematically attributes 30–40% of reported conversions via view-throughs that often credit the wrong channels during long sales cycles. These defaults were designed for short-cycle B2C e-commerce, not multi-stakeholder B2B sales.

LinkedIn’s default attribution window is 30-day click and 7-day view, which means roughly two-thirds of conversion events fall outside the default window for a company with an 84-day median sales cycle. An agency that leaves platform defaults unchanged is making budget decisions on a small slice of the data.

Questions to ask any agency:

• What attribution window are you using on each platform, and why?
• How does your window compare to our average sales cycle length?
• How do you handle closed-won events that fall outside the platform’s maximum window?

Red flags:

• Platform defaults left unchanged, such as 30-day click on Google or 7-day view on LinkedIn
• No process for matching CRM close dates to original ad clicks
• An attribution model that cannot explain which channels influenced pipeline created 90 or more days ago

Metric proving this criterion is met: companies using a 90-day multi-touch attribution window identify an average of 2.3 additional revenue-generating channels per quarter that a 30-day last-touch window had attributed to zero.

3. Primary-versus-Secondary Conversion Architecture That Trains on Revenue

The third criterion is conversion architecture, because the platform will chase whatever signal you feed it. The agency must separate the conversion events used for bidding optimization from those used only for reporting, or the algorithm will train itself on the wrong audience.

Modern ad platforms use machine learning to find more of whatever conversion event they receive. Pointed at a form fill, the algorithm finds the people most likely to fill out forms, such as students, competitors, and job seekers, while reporting a falling cost per conversion. Form fills, content downloads, newsletter signups, and unqualified chat initiations should be excluded from the primary conversion column once offline events are active, because their higher volume pulls Smart Bidding toward low-quality signals and inflates effective CAC.

The correct architecture designates SQL creation or opportunity creation as the primary conversion event, which becomes the signal the bidding algorithm optimizes toward, and tracks everything else as secondary. Secondary conversions remain visible in reporting but are never used for account-wide optimization. Making SQL the primary conversion event and switching to Target CPA bidding against it is the single highest-leverage change for B2B SaaS accounts generating 30 or more offline conversions per month.

SaaSHero maintains this architecture in every account as a condition of engagement, not as an optional configuration. Secondary conversions are tracked and visible, but they are never used for bidding.

Questions to ask any agency:

• What is your primary conversion event in our ad accounts right now?
• Which conversion events are excluded from account-wide optimization?
• How do you prevent form fills from influencing Smart Bidding?

Red flags:

• All conversion events weighted equally in the primary column
• No distinction between primary and secondary conversions in the account
• Optimization toward demo requests without downstream qualification data

Metric proving this criterion is met: SQL volume holds steady or increases while raw form-fill volume declines, which is the signature of an account trained on qualified outcomes instead of raw volume.

Learn how we configure your conversion architecture from day one to ensure the bidding algorithm trains on qualified outcomes rather than form fills.

4. CRM Sync and Offline-Conversion Imports That Close the Loop

The fourth criterion is CRM sync, because the algorithm cannot learn which clicks produce revenue until lifecycle events flow back into the ad platforms. The agency must connect CRM lifecycle events back to the ad platforms daily or the bidding algorithm will keep optimizing for surface-level leads.

Without offline conversion imports from CRM, Google Ads bidding algorithms optimize exclusively for visible form fills or demo requests, generating volume that appears productive in-platform but yields win rates less than half those of referrals or organic leads. The fix is to map the GCLID through at least the SQL stage and import closed-won revenue as an offline conversion event, which enables the algorithm to optimize toward actual deal value.

The missing offline conversion signal is the single biggest reason SaaS PPC programs optimize toward the wrong outcome instead of revenue, according to audits across 250+ B2B SaaS accounts. B2B SaaS accounts that implement CRM integration and offline conversion imports reach pipeline-coverage targets within two quarters 70% or more of the time, compared with 20% for accounts that skip the tracking and attribution phases.

B2B SaaS companies require 2–4 weeks of accumulated offline conversion data for Google Ads algorithms to begin reliable learning, with visible improvements in pipeline quality typically appearing within 30–60 days. This timeline means the integration must be built before launch, not retrofitted after the account has spent a quarter training on form fills.

Questions to ask any agency:

• How do you pass CRM lifecycle events back to Google Ads and LinkedIn?
• What is your offline conversion match rate, and how do you monitor it?
• How do you handle closed-won events that fall outside the 90-day GCLID window?

Red flags:

• No GCLID capture in the CRM
• Offline conversion match rate below 70%, which indicates CRM sync delays or GCLID capture failures
• Offline conversion import treated as a future phase rather than a launch requirement

Metric proving this criterion is met: offline conversion match rate above 70% and pipeline quality improving within 30–60 days of import activation.

5. Post-Click Ownership From Ad to CRM Record

The fifth criterion is post-click ownership, because accountability breaks the moment the agency stops owning the experience. The agency must own the landing page, including design, copy, build, hosting, and testing, or it cannot be held responsible for the outcome of the spend it manages.

Ad platform optimization determines who clicks, while landing page quality determines whether that click becomes a qualified lead. An agency that controls only the ad account is optimizing half the equation and reporting on the half it controls. The page the traffic lands on is the highest-leverage variable in the funnel, and headline copy is the single most impactful element on landing page conversion rate, yet it sits outside the scope of an agency that does not own the page.

The structural failure is organizational. The agency owns the ads, the client’s web team owns the page, the form belongs to marketing ops, and the conversion event belongs to whoever configured tag management. When performance drops, this fragmented ownership means no party is accountable, which turns diagnosis into a weeks-long exercise in finger-pointing. The only fix is to consolidate ownership so one party controls the full chain from impression to CRM record and can diagnose issues immediately.

SaaSHero designs, builds, hosts, and A/B tests every landing page its campaigns point to, using Figma for client approval and Unbounce for hosting and testing. Headline testing runs first because that is where the largest conversion leverage sits. The page ships without routing through the client’s web team backlog.

Questions to ask any agency:

• Do you build and host the landing pages your campaigns point to?
• Who writes the landing page copy and who designs the page?
• When was the last time you ran a headline A/B test on a client’s landing page?

Red flags:

• Agency delivers CRO recommendations for the client to implement
• Campaign traffic landing on the homepage or a generic product page
• No active A/B testing program on landing pages

Metric proving this criterion is met: a continuous A/B testing program running on purpose-built pages, with conversion rate improving quarter over quarter independent of traffic volume changes.

6. Proactive Strategy Cadence That Reduces Your Workload

The sixth criterion is strategy cadence, because a strong technical setup still fails if the agency waits for you to drive the roadmap. The agency must set the test agenda and arrive at every meeting with recommendations already made, not wait for the client to supply direction.

LTV/CAC and marketing-sourced revenue are displacing MQL volume as the headline metrics in CFO and board conversations. A marketing leader who cannot answer “what are we doing this month that we were not doing last month” is maintaining a program instead of running a proactive one. The test queue belongs to the agency, not the marketing leader.

A reactive agency relationship has a predictable signature. The client generates the ideas, assigns the work, chases the creative, and finds the problems in the account before the agency does. That pattern signals a scope failure, not just an execution failure, because the agency was hired to reduce the marketing leader’s workload and instead behaves like a direct report she cannot replace quickly.

A proactive cadence has fixed deliverables that arrive without being requested. These include monthly competitor analysis across paid search and paid social, quarterly budget reallocation analysis, continuous A/B testing with a documented rationale, and bi-weekly strategy calls where the agency presents what should change and why. Board-defensible reporting requires walking in with a pre-built reallocation trade-off ranked by marginal pipeline per dollar, which is only possible when the agency tracks that number continuously.

Questions to ask any agency:

• Who sets the test agenda, you or us?
• What standing deliverables do we receive without requesting them?
• How often do you run competitor analysis, and what does it cover?

Red flags:

• Strategy calls where the client supplies the agenda
• Competitor analysis delivered once at onboarding and never updated
• No documented testing program with a rationale for what is being tested and why

Metric proving this criterion is met: the client cannot remember the last time they told the agency what to test because the agency arrived with the proposal already written.

See how SaaSHero’s standing cadence of competitor analysis, budget reviews, and testing removes the marketing leader from the project-management seat.

7. 90-Day Validation Gate With Clean, Defensible Data

The seventh criterion is a 90-day validation gate, because you need a clear checkpoint where CAC and pipeline economics can be judged on real data instead of setup activity. The engagement must include a defined checkpoint at 90 days with enough clean data to evaluate channel economics, not just activity metrics.

B2B SaaS companies with $5M–$25M ARR had a median CAC payback of 15–18 months in 2026, and companies with $25M–$50M ARR had a median of 18–20 months, so a 90-day gate does not serve as a final verdict on payback. It does, however, provide the earliest point where the conversion architecture, messaging thesis, and channel structure can be evaluated on qualified pipeline instead of setup work.

A 90-day pilot commitment before moving to a longer engagement is the standard recommendation for staging financial risk on CAC payback. The gate serves two purposes. It gives the agency enough data volume for the bidding algorithms to learn from qualified signals, and it gives the buyer a defensible answer to the board question “is this working?”

The gate requires clean data from day one. An account launched on inherited tracking produces numbers nobody can defend at day 90 because the data set mixes old and new rules. Conversion tracking must be rebuilt at onboarding, not retrofitted at the gate, so the 90-day window contains clean, comparable data instead of a blend of inherited and corrected signals. Visible improvements in pipeline quality typically appear within 30–60 days of offline conversion import activation, so a properly configured account produces meaningful signal well before the 90-day gate and allows the checkpoint to evaluate established trends instead of early noise.

Questions to ask any agency:

• What specific metrics will we evaluate at 90 days, and how are they defined?
• Is conversion tracking rebuilt at onboarding or inherited from the previous setup?
• What happens if the 90-day data does not support continuing?

Red flags:

• No defined validation gate, only a rolling retainer with no checkpoint
• Inherited tracking carried forward without audit or rebuild
• 90-day review based on platform metrics rather than CRM-connected pipeline data

Metric proving this criterion is met: cost per SQL and pipeline coverage are measurable and defensible by day 90, with the data traceable to a clean conversion architecture and CRM sync that meet the match-rate threshold established earlier.

Frequently Asked Questions

How CAC Differs From Cost Per Lead When You Judge Agencies

Cost per lead (CPL) measures the spend required to generate a form submission or contact record, while CAC measures the total spend required to acquire a paying customer. In B2B SaaS, the two numbers can diverge dramatically. An agency that drives high lead volume at low CPL may simultaneously produce a CAC that exceeds 24 months of payback if those leads do not convert to qualified opportunities. Evaluating an agency on CPL rewards volume, while evaluating on CAC rewards revenue. The correct evaluation metric is cost per sales-qualified lead and cost per closed-won customer, connected to the CRM data that tracks what happened after the form was submitted.

How Agencies and RevOps Split CRM Integration Work

The agency must own the configuration of offline conversion imports and the primary-versus-secondary conversion architecture. RevOps owns the CRM itself, including lifecycle stage definitions, routing rules, and data hygiene, and serves as the essential internal ally for making the integration work. The practical division is that RevOps grants access and maintains the CRM, while the agency configures the GCLID capture, the import cadence, and the conversion event mapping. Without RevOps cooperation the integration cannot be built, and without the agency owning the configuration it will not be built correctly. SaaSHero treats RevOps as the most important internal stakeholder in any new engagement and involves them from the first week of onboarding.

How Long CAC Improvement Takes With a CRM-Connected Agency

The timeline for CAC improvement has two phases. The first phase covers setup, including conversion tracking rebuild, offline conversion imports activation, and the bidding algorithms accumulating clean data, which typically takes 30–60 days. The second phase covers optimization, where the algorithms have enough qualified signal to begin shifting spend toward higher-converting audiences and keywords. Meaningful improvement in pipeline quality, not just lead volume, typically appears within 60–90 days of a properly configured launch. CAC payback improvement is a lagging indicator that follows pipeline quality improvement by one full sales cycle, which for mid-market B2B SaaS is typically 60–120 days. A 90-day validation gate is the earliest defensible checkpoint for evaluating whether the channel economics are moving in the right direction.

Which Attribution Model Fits a $10M–$50M B2B SaaS Company

For most companies in this revenue range with sales-assisted funnels and 60–180 day sales cycles, a multi-touch model, either U-shaped (position-based) or W-shaped, provides a more accurate view than last-click. Last-click assigns full credit to the final touchpoint before conversion, which in a long B2B cycle is typically a branded search that occurred after the buying decision was already made. This pattern systematically defunds the upper-funnel channels that created the demand. The attribution window matters as much as the model, because a multi-touch model applied with a 30-day lookback window will still miss the majority of the customer journey for a company with an 84-day median sales cycle. The window should be set to at least 1.5× the average sales cycle length, with 90 days as the standard for mid-market and 180 days for enterprise or procurement-gated deals.

How SaaSHero’s Flat Retainer Works and What It Covers

SaaSHero charges a single flat monthly retainer indexed to total monthly ad spend under management. The retainer covers five capability areas, including paid media strategy and management across all major channels, creative work across concept, copy, and design, landing page design and build and A/B testing, attribution and CRM-connected reporting, and proactive strategy. There are no per-channel line items, no creative production fees, and no additional charges for adding or removing a channel. The Growth Team retainer starts at $4,000 per month and scales with total ad spend. Media spend is paid directly by the client to the platforms, and SaaSHero does not mark up media. All accounts, assets, design files, and dashboards are owned by the client throughout the engagement and transferred in full at offboarding.

Conclusion

The seven criteria in this scorecard progress from foundational commercial terms to advanced measurement gates for a clear reason. An agency with the wrong compensation model will give biased channel-mix advice regardless of how good its attribution is. An agency with the right compensation model but a 30-day attribution window will defund the channels that create demand. Each criterion acts as a prerequisite for the next. Start the evaluation with compensation model and attribution window, because those two criteria sort the market faster than any other question, then work through conversion architecture, CRM sync, post-click ownership, strategy cadence, and the 90-day gate.

SaaSHero is the only agency whose commercial model and measurement layer satisfy every row of this scorecard. A flat, spend-indexed retainer removes fee conflicts from channel-mix decisions. Full-chain CRM ownership, from conversion tracking rebuild through offline conversion imports through CRM-connected reporting, closes the loop between ad spend and closed-won revenue. In-house creative and landing page teams remove the post-click gap. A proactive strategy cadence removes the marketing leader from the project-management seat. A defined 90-day validation gate built on clean data from day one gives the board a defensible answer at the first checkpoint. Run your current account against this scorecard to identify gaps before committing to a new agency.