accessiBe logo
Growth Marketing Manager · Home Assignment

How I approach growth

A campaign strategy, data analysis and experimentation plan for accessiBe's Ontario AODA compliance campaign.

01 · The Campaign

Build urgency around a fixed compliance deadline

Ontario organizations with more than 20 employees must recertify their accessibility compliance every three years, and the current deadline falls on December 31, 2026. That deadline is the goal: build enough qualified pipeline for accessiBe's compliance offering before it arrives, without treating urgency as a substitute for qualification.

Month 1: Capture + Build
Month 2: Optimize + Nurture
Month 3: Convert + Accelerate
⚑ Dec 31, 2026
Regulation
AODA
Ontario orgs, 20+ employees, recertify every 3 years
Budget
Up to $50,000
Persona
Ops / HR
The people who own compliance recertification
Most B2B campaigns build a long funnel: warm the audience with awareness and consideration content first, then push down funnel toward conversion over months. With a fixed December 2026 deadline, we don't have that runway. Instead we're running an accelerated funnel: starting directly with conversion campaigns from week one, while building the retargeting pool at the same time, and investing in it as it matures.
02 · Campaign Strategy & Budget

Three phases, one budget, rebalanced by performance

Initial allocation, rebalanced based on performance.

Month 1: Capture + Build

Capture existing demand, seed the rest

  • High-intent Google Search live from day one
  • LinkedIn prospecting + Lead Gen Forms
  • Meta Advantage+ audience prospecting
  • Awareness & video, build demand and future retargeting pools
  • Start building retargeting pools
Month 2: Optimize + Nurture

Follow the signal, not the plan

  • Shift budget to the winning channel × audience × creative
  • Increase retargeting weight
  • Keep high-intent Search funded
  • Sharpen compliance / deadline messaging
  • Lean on engagement & first-party audiences
Month 3: Convert + Accelerate

Push everything toward the deadline

  • Heavy retargeting on warm pools
  • High-intent Search sustained
  • Strong deadline / urgency messaging
  • Lead Gen + Demo campaigns
  • Faster lead routing, faster meeting booking
ChannelRoleMonth 1Month 2Month 3TotalShare
GoogleBrand, competitor & high-intent search: demand capture$3,500$3,000$3,000$9,50019%
LinkedInPrecision Ops/HR targeting: qualified demand generation$7,000$8,000$8,500$23,50047%
MetaBroaden reach beyond LinkedIn's precision targeting; convert via lookalikes & retargeting$5,500$6,000$5,500$17,00034%
Total$16,000$17,000$17,000$50,000100%

The 47% LinkedIn allocation is an initial starting allocation, not a fixed commitment.

Google

Demand capture, including brand + competitor. High-intent search around AODA, accessibility compliance, compliance deadlines, and defensive brand/competitor terms. Kept deliberately lean: in B2B, absolute search volume for a compliance term like this is low, so this channel protects intent rather than driving scale.

LinkedIn

Reach the right professionals. Job-function targeting, Lead Gen Forms, website conversion, thought-leadership video, retargeting. Gets the largest share of budget because it is the most precise B2B channel available for this persona.

Meta

Extend reach, then convert. LinkedIn's targeting is precise but narrow; Meta's Advantage+ audience (Meta's own automated, algorithmic targeting) reaches people LinkedIn's filters miss, then MQL lookalikes and retargeting bring the qualified ones back to convert.

03 · Audience Strategy

LinkedIn identifies the right professionals. Meta identifies the right people.

On LinkedIn specifically, we are running two distinct audiences with two distinct messages, not one merged Ops/HR list.

LinkedIn · audience 1

HR trackCompliance as an HR obligation

CHROVP HRHead of HR Sr. Director HRRegional HR DirectorHR DirectorCEO (small orgs)

Message angle: AODA recertification is now an HR compliance obligation, not a legal footnote. Position accessiBe as the fastest path to recertifying on time.

LinkedIn · audience 2

Ops trackCompliance as an operational risk

COOVP OperationsHead of Operations Sr. Director OperationsDirector of OperationsCEO (small orgs)

Message angle: recertification is an operational deadline with real penalties for missing it. Position accessiBe as the lowest-risk way to hit it.

Targeting the ICP within platform constraints
  • 11+ employees is the closest practical LinkedIn company-size filter to the legal >20-employee threshold, not an exact match.
  • Smaller organizations may not have dedicated HR/Ops leadership, which is why CEO is included on both tracks.
  • Qualification still happens downstream: managers and too-small companies are filtered out at the lead-qualification stage, not the targeting stage.
Meta · audience
  • Advantage+ audience: Meta's own automated, algorithmic targeting, running as always-on prospecting
  • MQL lookalike: seeded from accessiBe's actual marketing-qualified leads, running in parallel
  • Retargeting: website visitors, relevant page visitors, video/social engagers, running in parallel

Future: an SQL / customer lookalike, once enough closed-won data exists.

Don't teach Meta what an HR manager looks like. Teach Meta what a good lead looks like.
04 · User Journey & Measurement

Optimize toward pipeline, not lead volume

Ad / Search / Video Landing Page / Lead Gen Form Lead MQLChili Piper Meeting SQL Pipeline / Revenue
Retargeting loop, running underneath the whole journey: Ad / Video / Website Interaction → Retargeting → Conversion.

Primary · Business

Whether we're building revenue-worthy opportunities, not just activity.
Metrics SQLsPipeline generatedCost per SQLPipeline efficiency
Measure on Cohort-based (SQL)Monthly / campaign-to-date (pipeline)Longer term (revenue)

Secondary · Quality

Whether the leads we're generating are the right leads: audience and message fit.
Metrics MQL volumeMQL rateCost per MQLMQL → SQL
Measure on 2 to 3 weeks (MQL rate, CPMQL)Cohort-based (MQL → SQL)

Diagnostic

Whether the mechanical top of funnel is working.
Core conversion LeadsLP CVRCPLDemo bookingsMeeting rate
Engagement / delivery ClicksCTRCPC
Measure on Daily / weekly
05 · When Performance Is Off Track

Find the broken stage before changing the strategy

Three weeks in, performance is below expectations. First identify where the funnel is breaking, because each broken stage requires a different fix.

Not enough SQL
MQLs are coming in, but too few are becoming sales-qualified.

Possible issue: sales follow-up, routing, or lead readiness.

Fix: speed up lead routing and follow-up. If it's readiness, shift the message toward deadline-aware, high-intent buyers, for example: "Recertification is due December 2026, here's the fastest compliant path."
Not enough MQLs
Leads are coming in, but too few meet the qualification criteria.

Possible issue: persona, targeting, offer, or message.

Fix: tighten targeting, change the offer, retest messaging against the actual persona.
Not enough leads
Traffic is reaching the funnel, but too few visitors become leads.

Possible issue: landing-page behavior, form friction, or CTA.

Fix: optimize the behavior first, not the media buy, landing page, form length, CTA.
06 · Campaign Dashboard

The data shows where quality is being won, and lost

452 ad sets · 5 ecosystems · 2 regions

Spend
$302,986
452 ad sets
Clicks
75,463
CTR 0.53%
Conversions
3,163
CVR 4.19% of clicks
MQL
41
MQL rate 1.30%
CPMQL (blended)
$7,390
CPL $95.79

Cost per MQL, by ecosystem

Lower is better. Widget shown separately: it produced zero MQLs, so cost per MQL is undefined, not "free."
wcag
$1,965
MQL rate 5.03% · $17.7K spend · 9 MQL
ada
$6,062
MQL rate 1.41% · $66.7K spend · 11 MQL
accessibility
$9,844
MQL rate 1.34% · $39.4K spend · 4 MQL
scan
$9,950
MQL rate 0.98% · $169.2K spend · 17 MQL
widget
0 MQL
$59.32 CPL, cheapest in the dataset · $10.1K spend · 0 MQL
$0$2.5K$5K$7.5K$10K

Spend share, by ecosystem

Differences in totals come from ad-set count, not per-set budget.
scan 55.8% · $169.2K
ada 22.0% · $66.7K
accessibility 13.0% · $39.4K
wcag 5.8% · $17.7K
widget 3.3% · $10.1K
EcosystemAd setsSpendShareClicksCTRCPCConv.CVRCPLMQLMQL%CPMQL
scan246$169,15155.8%36,5050.53%$4.631,7354.75%$97.49170.98%$9,950
ada108$66,68722.0%23,4790.52%$2.847813.33%$85.39111.41%$6,062
accessibility58$39,37413.0%8,4000.55%$4.692983.55%$132.1341.34%$9,844
wcag24$17,6895.8%3,3660.58%$5.251795.32%$98.8295.03%$1,965
widget16$10,0853.3%3,7130.51%$2.721704.58%$59.3200.00%n/a

NAMER

Ad sets248
Spend$175,823
Clicks33,182
CVR (clicks)5.31%
MQL26
MQL rate1.48%
CPMQL$6,762
CPC$5.30

EMEA

Ad sets204
Spend$127,163
Clicks42,281
CVR (clicks)3.31%
MQL15
MQL rate1.07%
CPMQL$8,478
CPC$3.01
NAMER wins on conversion volume, MQL volume, MQL rate and cost per MQL. EMEA's only real edge is cheaper media cost (CPC), and cheaper clicks alone don't carry through to better MQL economics here.
Ecosystem × RegionAd setsSpendClicksConv.CVR (LP)CPLMQLMQL rateCPMQL
wcag × namer20$15,0302,5931568.67%$96.349
5.77%
$1,670
ada × emea93$57,34821,5936894.98%$83.2311
1.60%
$5,213
accessibility × namer27$18,9563,9591325.92%$143.603
2.27%
$6,319
scan × namer186$132,50024,7441,3827.34%$95.8814
1.01%
$9,464
scan × emea60$36,65111,7613534.74%$103.833
0.85%
$12,217
accessibility × emea31$20,4184,4411664.91%$123.001
0.60%
$20,418
ada × namer15$9,3391,886926.74%$101.510
0.00%
n/a
wcag × emea4$2,660773234.94%$115.650
0.00%
n/a
widget × emea16$10,0853,7131706.97%$59.320
0.00%
n/a
Strongest quality signal

WCAG × NAMER. Small, but the clearest proof of what a well-matched audience and message can do.

$15.0K spend8.67% LP CVR5.77% MQL rate$1,670 CPMQL
Biggest scale combination

Scan × NAMER drives the most volume, but its downstream efficiency is weak.

1,382 of 3,163 total conversions$9,464 CPMQL
07 · From Data to Action

Turn the data into specific optimization moves

01

High volume, weak downstream quality

Scan × NAMER
24.7K clicks1,382 conversions14 MQLs$132.5K spend 1.01% MQL rate$9,464 CPMQL
Audit and tighten audience / keyword intent, then test new messaging and landing-page alignment. Goal: more MQLs from existing traffic and lower CPMQL before scaling spend.
02

Learn from the strongest quality signal

WCAG × NAMER
2.6K clicks156 conversions9 MQLs 5.77% MQL rate$1,670 CPMQL
Identify what's driving the quality, audience, keyword intent and messaging, and apply those learnings to other activity. Then test whether performance holds as volume increases.
03

Already working, worth sharpening before scaling

ADA × EMEA
21.6K clicks689 conversions11 MQLs 1.60% MQL rate$5,213 CPMQL
Unlike #01, this segment already beats the account's blended MQL rate. Refine targeting, keyword intent and ad-to-LP message match, then test a more relevant offer to push MQL rate and CPMQL further before adding budget.
04

Rebuild, don't optimize for cheap conversions

Widget
3.7K clicks170 conversions0 MQLs$10.1K spend$59 CPL
Change the optimization event from conversion volume toward qualified leads. Reassess the audience, offer and conversion path to identify why none of the conversions become MQLs.
08 · A/B Testing

Test the message. Measure the quality.

Choose the variant that produces higher-quality demand, not simply cheaper leads.

Test 1: value proposition, ease vs. risk

A
accessiBe homepage, variant A: Make Your Website ADA Compliant in Minutes
B
accessiBe homepage, variant B: Reduce the Risk of Costly Accessibility Claims
Hypothesis
Are prospects moved more by ease/speed, or by risk/pain?
Primary metric
Landing page CVR
Quality guardrail
MQL rate / CPMQL
Decision logic
If the higher-CVR headline also drags MQL rate down, we don't ship it, we'd rather keep the version that brings in fewer, better-fit leads.

Test 2: CTA / offer, audit vs. trial

A
accessiBe homepage, variant A: Get a Free Audit
B
accessiBe homepage, variant B: Start Free Trial
Hypothesis
Audit lowers friction for problem-aware visitors; trial filters for higher intent
Primary metric
MQL rate / MQL volume
Secondary
LP CVR, CPL, CPMQL, then SQL rate once cohorts mature
Decision logic
Pick the variant with the higher MQL rate, not the lower CPL, then confirm with SQL rate once each cohort has had time to convert.
09 · Closing

A loop, not a one-time launch

1Build 2Measure 3Diagnose 4Optimize 5Experiment 6Scale

Three separate questions, one working style: start from the outcome we actually need, let the data show where to lean in, and test our way to the rest.

accessiBe logo

Thank you

Dana Ephrat · dana.ephrat@gmail.com