Three years ago, a performance team built audiences by hand in the META Ads Manager: interests, lookalikes, exclusions, one ad set per segment. Today META makes Advantage+ the default. Most targeting inputs only count as suggestions. The META Ads AI now decides who sees an ad. A marketing lead therefore has to take care of two things: enough distinct creatives and a feedback loop from the CRM (customer relationship management) system.
Key points
- Since February 2025, Advantage+ has been the default for Sales, App and Leads campaigns. Only location, minimum age, language and excluded Custom Audiences remain hard limits[1].
- Interests, lookalikes, age and gender are suggestions. META may deliver beyond them[2].
- The creative takes over part of audience selection. META recommends many creatives with different themes and messages[3].
- The AI optimizes for the event you give it. If you only report “lead”, you get lots of leads. If you report SQLs (sales qualified leads, meaning leads the sales team has qualified) back from your CRM, you give it a better target.
- Measurement has narrowed: shorter view windows, and only a link click still counts as a click[4, 5].
Eight changes between December 2024 and fall 2026. More dates are in the text.
Before: the team built the audience
Until about 2024, work in the Ads Manager looked like this: for Sales campaigns, the team chose between a manual campaign and Advantage+ Shopping; for App campaigns, there was Advantage+ App. Leads campaigns were set up by hand[1]. The team created a separate ad set for each audience, for example interests, lookalikes and retargeting. It excluded interests, picked placements manually and set a budget cap for existing customers.
The logic: if you describe the audience precisely, you control who sees the ad. A good media buyer knew the interest categories and tested segment against segment.
Until about 2024, the team defined the audience field by field: one ad set per segment, with interests, age, exclusions and placements as fixed limits.
What changed with META Ads AI between 2024 and 2026
Advantage+ becomes the default
In February 2025, META began to simplify campaign setup. Since then, Sales, App and Leads campaigns no longer offer a choice between “manual” and “Advantage+”. Advantage+ Shopping became Advantage+ Sales[6]. Advantage+ audience, placements and budget are preset. If you change them, the “Advantage+ on” label switches off[1]. Since October 2025, META has been phasing out the legacy APIs for Shopping and App campaigns[7].
Several AI models control delivery
Several models work behind ad delivery. Andromeda selects a few thousand candidates for one person from tens of millions of ads[8]. GEM, META’s Generative Ads Model, passes its knowledge on to smaller models that score and rank ads[9]. On its third-quarter 2026 earnings call, META reported further conversion gains from GEM and Andromeda[10].
Targeting becomes a suggestion
The targeting options still exist. But META only treats location, minimum age (up to 25), languages and excluded Custom Audiences as hard limits. Maximum age, gender, Custom Audiences and Detailed Targeting are suggestions[1]. For eleven common performance goals, including leads and conversions, the suggestion feature for Detailed Targeting cannot be switched off[2]. META phased out exclusions based on interests and behaviors in 2024/25. Since January 2025, no new exclusions of this kind can be created. Existing ad sets kept running until someone edited them[11]. META has merged or removed many interests. Since January 15, 2026, ad sets that use removed options no longer deliver[12].
Placements are also hard to control now. Since fall 2026, META has been removing placement selection at the ad set level for campaigns with the Sales, Leads, App Promotion and Engagement objectives. Jon Loomer sums it up: “Meta is eliminating placement control at the ad set level.” Placements can then only be excluded at the account level, per campaign objective. The team can also use placement value rules to lower bids for individual placements. Awareness and Traffic campaigns keep placement selection for now[13, 14]. The sources give no fixed date for all accounts.
Measurement narrows
Since January 12, 2026, the Ads Insights API no longer returns data for 7-day and 28-day view windows[4]. Since March 2026, only a click on the ad’s link counts as a click. Other interactions fall under the new “engage-through”. The reporting default is now 7-day click, 1-day engage-through and 1-day view[5]. If you compare 2025 figures with 2026 figures, you are comparing different definitions.
A special rule applies in the EU
Following proceedings under the Digital Markets Act, META committed to giving users in the EU a choice from January 2026. They can share all their data, or share less data and see less personalized ads[15]. For the earlier version of these ads, META cited “70% fewer onsite conversions and 61% fewer offsite conversions”[16]. These figures come from META itself, from an ongoing legal dispute. There is no public data on how many people in Germany choose the restricted version.
What this means for marketing leads (assessment based on the sources above):
- Part of your reach runs on less data. For these people, the AI has fewer signals to learn from. Expect results in the EU to fluctuate more.
- Your own data carries more weight. Feedback from your CRM with valid consent, for example via the Conversions API, is data your team controls. Keep the consent rate in your cookie banner on your KPI list.
- Do not compare figures before and after January 2026 directly. If conversions in the Ads Manager have dropped since then, check your CRM figures first. That tells you whether performance changed or only the measurement.
- Align expectations early. Agree with management and sales which metric counts: cost per SQL or per purchase from the CRM. The number of leads in the Ads Manager is not enough as a yardstick.
Today only location, minimum age, language and excluded Custom Audiences are hard limits. Everything else is a suggestion. You steer through the performance goal and the creative.
Then and now
| Area in the Ads Manager | Then (until about 2024) | Now (as of November 2026) |
|---|---|---|
| Campaign setup | Manual; for Sales, Advantage+ Shopping as an alternative | Advantage+ preset; Advantage+ Sales replaces Shopping[6] |
| Audience | Interests, lookalikes, age, gender as fixed limits | Hard limits only for location, minimum age, language, excluded Custom Audiences; everything else is a suggestion[2] |
| Exclusions | Interests and behaviors could be excluded | Custom Audiences only[12] |
| Structure | One ad set per segment | A few large campaigns; up to 150 ads in new Advantage+ Sales campaigns[1] |
| Placements | Chosen by hand, per ad set | For Sales, Leads, App Promotion and Engagement, no selection in the ad set; only account-level exclusions per campaign objective and placement value rules[13, 14] |
| Creative | A few ads per segment | Many distinct creatives; AI text, AI images and video from still images[17, 18] |
| Control | Own rules and reports | Opportunity Score (0–100) counts accepted recommendations; it does not measure performance[19] |
| Attribution | Default: 7-day click (any click on the ad), 1-day view; 28-day click only existed in earlier years | 7-day click (link clicks only), 1-day engage-through, 1-day view[5] |
Pros and cons
Pros:
- Less manual setup. Audience, placements and budget are preset[1]. The team gains time for creative and testing.
- Measurable gains, according to META. META says GEM delivered 5% more conversions on Instagram and 3% more in Facebook Feed in Q2[9]. In META’s test of Advantage+ for leads, cost per qualified lead fell by 10%[6].
- New audiences. META writes that with many creatives on different themes, the AI can also find audiences a team did not expect[3].
Cons:
- Decisions are hard to trace. According to Digiday, agencies describe the system as a “black box”; clients are “curious, but cautious”[20].
- The success figures come from the vendor. Almost all of the reported gains come from META. META has not published a methodology for its AI assistance figures[21]. Independent tests reach different results. The measurement company Haus analyzed 640 incrementality tests: Advantage+ outperformed manually managed campaigns in only 42% of cases. Haus itself sells measurement solutions[22].
- Limits have gaps. According to Loomer, exclusions do not apply completely, and location targeting is imprecise[2]. The budget cap for existing customers no longer exists (see box)[1].
- Less comparability. Shorter windows and the new click definition break time series[4, 5].
- The AI only knows its target. It optimizes for the event set in the Ads Manager. It only sees what happens later in sales if someone reports it back. The section after next covers this.
Note
When retargeting looks like new customer acquisition
In Advantage+ Shopping, you could set a maximum share of budget for existing customers. Advantage+ Sales no longer has this cap[1]. The AI spends the budget where conversions are cheapest. These are often people who were going to buy anyway: existing customers, cart abandoners, returning visitors. The report shows these purchases as campaign results. It does not show whether the ad caused them.
What you can do:
- Upload existing customers as a Custom Audience and exclude them. META treats Custom Audience exclusions as a hard limit, though according to Loomer not without gaps[2].
- Manage new and existing customers separately with the Customer Lifecycle Strategy[18].
- Run an incrementality test to check how many additional purchases the ads bring.
The creative takes over part of the targeting
If targeting is only a suggestion, the AI selects people based on other signals. One of them is the ad itself. META built Andromeda to pick the right candidates for each person from many ad variants[8]. META defines creative diversification as “the practice of creating a wide range of ad creatives” for different audiences[3]. Jon Loomer recommends creating ads for different people and situations: different pain points, different formats, different customer types. In his view, small changes to one idea are not enough[23].
META does not give an official minimum number of creatives. Practitioners also give very different numbers. The agency Precis recommends starting with three concepts[24]. Loomer observes that 10, 20 or more ads per ad set have become common, and writes: “But more isn’t necessarily better.”[25] In a guest article for AdExchanger, Shukla of the agency Mile Marker warns that too many ads on a small budget spread it so thin that no single ad collects enough data[26]. So match the number of ads to the budget. For a B2B budget, three to five clearly different concepts are a good start. The variants must differ clearly. Five ads with the same message and a different background color give the AI little to choose from.
Example: five creatives for one offer
The following example is fictional. Product, figures and copy are invented for illustration only.
Offer: “Taktwerk”, shift-planning software for manufacturers with 50 to 500 employees. Goal: demo requests. Each of the five ads differs in exactly one clear dimension.
| # | Dimension | Headline | Image idea | What the ad “tells” the AI |
|---|---|---|---|---|
| 1 | Problem | “Friday, 4 p.m.: Two sick calls, and Monday’s schedule is back to square one.” | Paper shift schedule full of crossed-out entries | Reaches people with acute scheduling problems, usually shift supervisors |
| 2 | Role | “For plant managers: labor cost per shift at a glance.” | Dashboard with three KPIs | Reaches decision-makers with budget responsibility |
| 3 | Proof / figure | “20 minutes instead of 4 hours: how a 180-employee plant plans its shifts.” (fictional figure) | Before-and-after clock | Reaches people who want proof before they decide |
| 4 | Format | “See how the schedule works in 30 seconds.” | Short screen recording of the software | Reaches people who want to see the product before they get in touch |
| 5 | Emotional instead of rational | “Home on time again.” | Shift supervisor leaves the plant at 5 p.m. | Reaches people through the personal outcome, without figures |
The right-hand column is an assessment. META does not publish which features of a creative its models evaluate. The assumption behind it: if different people respond to ad 1 than to ad 2, the system learns whom to show which ad. If all five are the same, there is nothing to tell apart. But the system does not say why an ad works[26]. So decide before launch what each variant is meant to test.
Fictional example: five ads for the invented software “Taktwerk”. Each one differs in exactly one dimension.
For marketing leads, this means: creative is now also audience work. Whoever sets the five dimensions helps decide which people the campaign reaches. That decision belongs in campaign planning.
The risk: the AI optimizes for the wrong target
The AI does the job it is given. If the optimization goal is “lead”, it looks for people who fill in a form. It does not know whether they buy later. A lead form inside META (instant form) is quick to fill in. That lowers the cost per lead. These leads are not automatically worse. In a test by Jon Loomer, the share of good leads was almost the same for META lead forms and website forms: 29.1% versus 29.0%. Loomer stresses that a single test does not prove this[27]. The result depends on the goal the campaign optimizes for and on how the form is built.
The form filters as well. One to three qualifying questions, a review step before submitting and a required business email address lower the number of leads and raise the share of good ones. The B2B agency Directive recommends this: “Every added field needs to change a decision.”[28] Loomer puts it more briefly: “If you want higher quality leads, ask more questions.”[27]
In my article on LinkedIn Ads for B2B SaaS, I looked at measuring cost per SQL instead of cost per lead. META adds one step. There, an AI controls delivery. The metric in the report is no longer enough. The AI itself has to know the better target.
Lead, MQL, SQL, customer: the stages in the CRM
HubSpot defines the default stages as follows[29]:
- Lead: a contact who converted on the website or through another touchpoint.
- Marketing Qualified Lead (MQL): “a contact or company that your marketing team has qualified as ready for the sales team”.
- Sales Qualified Lead (SQL): “a contact or company that your sales team has qualified as a potential customer”.
- Opportunity: a contact associated with a deal.
- Customer: “a contact or company with at least one closed deal”. In HubSpot’s default pipeline, the matching deal stage is called “Closed won”[30].
For leadership: a person decides two of these transitions. The marketing team decides on MQL, sales on SQL. These decisions are the signal the AI is missing.
Fictional calculation for illustration: Campaign A delivers leads at 40 €, and 5% of them become SQLs. One SQL therefore costs 800 €. Campaign B delivers leads at 90 €, and 20% of them become SQLs. One SQL costs 450 €. If you only look at cost per lead, you scale up campaign A. An AI that only receives leads does the same.
| A: cheap leads | B: targeted leads | |
|---|---|---|
| Monthly budget (fictional) | 7,200 € | 7,200 € |
| Cost per lead (CPL) | 40 € | 90 € |
| Leads per month | 180 | 80 |
| SQL rate | 5% | 20% |
| SQLs per month | 9 | 16 |
| Cost per SQL | 800 € | 450 € |
The formula: cost per SQL = CPL ÷ SQL rate. At 40 € CPL and a 5% SQL rate, that is 40 ÷ 0.05 = 800 €.
How the stages get back to META
META offers the Conversions API for this. You can also use it to upload events from the CRM, such as stages in the sales process[31]. META planned to discontinue the earlier Offline Conversions API in May 2025. For new integrations, META recommends the Conversions API[32].
In a Leads campaign, you choose between two performance goals. According to META, “Maximize number of leads” optimizes for volume, meaning as many completed forms as possible[33]. “Maximize number of conversion leads” optimizes for quality: your CRM reports back which leads reach a later stage, and META learns which leads are most likely to convert[33, 31]. This goal is currently only available for instant forms[33].
Option 1: Conversion leads with META lead forms. The conversion leads performance goal optimizes for a later stage in the CRM. META’s documentation is inconsistent here. The main page only mentions lead ads with instant forms[31]. The FAQ for the same integration mentions lead ads with instant forms or website forms[34]. On the main page, META lists these requirements[31]:
- at least 200 leads per month;
- data uploaded at least once a day;
- the optimized stage occurs within 28 days of the lead;
- the conversion rate of this stage is between 1% and 40%;
- about one month until full effect, including a learning phase of two to four weeks.
In the FAQ, META gives one to two months of training time[34]. Plan for two months.
The CRM should report every stage as soon as it changes, including the first lead stage[35]. In the Ads Manager, the team selects the earliest stage the AI should optimize for. According to META, the system then also optimizes for all later stages and may choose a different stage[36].
Option 2: HubSpot sends lifecycle stages to META. HubSpot can send lifecycle stage changes and form submissions to META via the Conversions API. This is available from Marketing Hub Starter upward. Starter allows up to 5 such events, Professional 50, Enterprise 100[37]. The team should know four points from the HubSpot documentation:
- HubSpot only counts changes made after the event was created. It does not send older data.
- HubSpot only syncs forms that were created in HubSpot.
- META only supports lifecycle data from contacts who interacted with lead ads campaigns. For optimization, the campaign must use the conversion leads performance goal.
- HubSpot recommends installing the META pixel only through HubSpot, so that events do not arrive twice.
If you collect leads through forms on your own website, you can also optimize for SQLs. The conversion leads performance goal is not available for this. The route is standard optimization for website conversions:
- In the ad set, select “Website” as the conversion location.
- Connect the CRM to META via the Conversions API. When sales qualifies a lead as an SQL, the CRM sends a custom event to META, for example “qualified lead” or a value per lead. For META to match the event to the original click, the event must include customer information parameters such as the click ID (
fbc) and the browser ID (fbp)[38]. - Select this event as the optimization goal in the ad set.
META describes the basis for this: it processes server events from the Conversions API like pixel events, and they “may be used in measurement, reporting, or optimization”[39]. The conversion leads performance goal, on the other hand, is according to META “currently only compatible with Facebook/Instagram’s Lead Ads (Instant Forms)”[31]. The B2B SaaS agency Kalungi also describes the route via a custom CRM event[40]. This route needs enough cases per week. The agency Blck Alpaca cites META’s rule of thumb of about 50 results per ad set per week[41]. If you get fewer SQLs, optimize for an earlier stage such as MQL and steer quality through the questions in the form.
The CRM reports every stage back to META via the Conversions API. The campaign optimizes for the stage that signals revenue, such as the SQL.
Who is responsible for what
- Marketing lead: sets the optimization goal (e.g. SQL instead of lead) and the metric in the report (cost per SQL). Approves which data goes to META.
- Sales: defines what an SQL is and updates the status promptly. If sales only updates the status after six weeks, the stage falls outside the 28-day window. Sales also responds quickly to new inquiries. For demo or pricing requests, the B2B agency Directive recommends a first personal response within 15 minutes[28]. Lead quality is also decided after the form[42].
- CRM / RevOps (revenue operations): maps the stages to events, checks data quality and duplicates, and agrees with the people responsible for data protection which contacts are transferred. HubSpot offers a choice between all contacts in a stage and only contacts who interacted with ads[37].
- Performance team: builds the campaigns, supplies the creative variants and sets up the events in the Ads Manager.
What to do now
- Check the optimization goal. Look at every Leads campaign: which event does it optimize for? If it only says “lead”, start there.
- Define the stages. Agree in writing with sales what an MQL and an SQL are. Check whether the SQL stage is usually reached within 28 days. Check the volume too: conversion leads requires at least 200 leads per month[31]. If your account does not reach that, optimize for an earlier stage such as MQL first[40].
- Set up the feedback loop. With lead ads: conversion leads and the CRM integration. With HubSpot: lifecycle events via the Conversions API[37]. Allow one to two months until full effect[31, 34].
- Plan the creative. Several variants per campaign, each differing in one clear dimension: problem, role, proof, format, tone.
- Consolidate campaigns. A few large campaigns per objective. Separate existing customers with excluded Custom Audiences[2].
- Keep your own measurement. Reconcile the CRM figures with the Ads Manager. Re-baseline time series around January and March 2026. Use the Opportunity Score as a checklist; it does not measure performance[19]. With larger budgets, run a lift test (an incrementality test) to check how many additional deals the ads bring[5, 22].
- Track the EU effect. Compare conversions since January 2026 with closed deals in the CRM.
Conclusion
The days when a team built the audience in the Ads Manager segment by segment are largely over. Today, the AI selects who sees an ad. The team still controls two things above all: the range of creatives and the goal the AI optimizes for. Then there are the offer, the form and the sales follow-up[24, 41, 42]. These are leadership decisions. Without a feedback loop from the CRM, the AI only learns who fills in forms.
Further reading: How to measure cost per SQL instead of cost per lead is covered in B2B SaaS LinkedIn Ads – 3 Pitfalls & The 2026 Strategy Guide. How a team plans content from its business goals is described in the free Content Engine Playbook (PDF).
Sources
- Loomer, J. (2025) ‘Advantage+ Sales, App, and Leads Campaigns are Coming’. Jon Loomer Digital, 10 February 2025. Available at: https://www.jonloomer.com/advantage-plus-sales-app-leads-campaigns/ (Accessed: 17 November 2026).
- Loomer, J. (2026) ‘When Your Targeting Inputs Are Only Suggestions’. Jon Loomer Digital, 16 February 2026 (updated 2 April 2026). Available at: https://www.jonloomer.com/targeting-suggestions/ (Accessed: 17 November 2026).
- Meta (2025) ‘The Creative Advantage: Unlocking the Power of Diversification with Meta Andromeda’. Meta for Business, 22 April 2025. Available at: https://www.facebook.com/business/news/the-creative-advantage-unlocking-the-power-of-diversification-with-meta-andromeda (Accessed: 17 November 2026).
- Rijo, L. (2025) ‘Meta restricts attribution windows and data retention in Ads Insights API’. PPC Land, 18 October 2025. Available at: https://ppc.land/meta-restricts-attribution-windows-and-data-retention-in-ads-insights-api/ (Accessed: 17 November 2026).
- Loomer, J. (2026) ‘How Meta Ads Attribution Works in 2026’. Jon Loomer Digital, 10 March 2026. Available at: https://www.jonloomer.com/meta-ads-attribution-2026/ (Accessed: 17 November 2026).
- Hutchinson, A. (2025) ‘Meta Rolls Out Updates to Advantage+ AI Ad Targeting Tools’. Social Media Today, 6 February 2025. Available at: https://www.socialmediatoday.com/news/meta-advantage-plus-ad-updates-february-2025/739472/ (Accessed: 17 November 2026).
- Rijo, L. (2025) ‘Meta deprecates legacy campaign APIs for Advantage+ structure’. PPC Land, 26 October 2025. Available at: https://ppc.land/meta-deprecates-legacy-campaign-apis-for-advantage-structure/ (Accessed: 17 November 2026).
- Sun, C., Yu, N., Lu, H., Wang, L., Pu, Y., Liu, G., Bhatia, N. and Musumeci, G. (2024) ‘Meta Andromeda: Supercharging Advantage+ automation with the next-gen personalized ads retrieval engine’. Engineering at Meta, 2 December 2024. Available at: https://engineering.fb.com/2024/12/02/production-engineering/meta-andromeda-advantage-automation-next-gen-personalized-ads-retrieval-engine/ (Accessed: 17 November 2026).
- Li, H., Liu, X., Nie, J., Wen, E., Yang, C., Yang, J., Yu, N., Beg, H., Arditi, G. and Bhatia, N. (2025) ‘Meta’s Generative Ads Model (GEM): The Central Brain Accelerating Ads Recommendation AI Innovation’. Engineering at Meta, 10 November 2025. Available at: https://engineering.fb.com/2025/11/10/ml-applications/metas-generative-ads-model-gem-the-central-brain-accelerating-ads-recommendation-ai-innovation/ (Accessed: 17 November 2026).
- Meta Platforms (2026) ‘Meta Platforms, Inc. (META) Third Quarter 2026 Results Conference Call’ [Transcript, PDF]. Meta Investor Relations, 21 October 2026. Available at: https://s21.q4cdn.com/399680738/files/content_files/META-Q3-2026-Earnings-Call-Transcript.pdf (Accessed: 17 November 2026).
- Alston, N. and Hakobyan, R. (2025) ‘Removal of Detailed Targeting Exclusions’. Meta for Developers Blog, 21 January 2025. Available at: https://developers.facebook.com/blog/post/2025/01/21/removal-of-detailed-targeting-exclusions/ (Accessed: 17 November 2026).
- Hutchinson, A. (2025) ‘Meta Is Consolidating More of Its Detailed Ad Targeting Options’. Social Media Today, 17 August 2025. Available at: https://www.socialmediatoday.com/news/meta-removes-more-detailed-ad-targeting-options-facebook-instagram/757856/ (Accessed: 17 November 2026).
- Loomer, J. (2026) ‘Meta Clarifies Upcoming Placement Control Changes’. Jon Loomer Digital, 27 October 2026. Available at: https://www.jonloomer.com/meta-placement-control-changes/ (Accessed: 17 November 2026).
- Hutchinson, A. (2026) ‘Meta offers updated ad placement control guidance’. Social Media Today, 12 October 2026. Available at: https://www.socialmediatoday.com/news/meta-offers-updated-ad-placement-control-guidance/832791/ (Accessed: 17 November 2026).
- European Commission (Directorate-General for Competition; Directorate-General for Communications Networks, Content and Technology) (2025) ‘Meta commits to give EU users choice on personalised ads under Digital Markets Act’. Digital Markets Act, 8 December 2025. Available at: https://digital-markets-act.ec.europa.eu/meta-commits-give-eu-users-choice-personalised-ads-under-digital-markets-act-2025-12-08_en (Accessed: 17 November 2026).
- Lamb, T. (2025) ‘Why the Commission’s Decision Undermines the Goals of the DMA’. Meta Newsroom, 2 July 2025. Available at: https://about.fb.com/news/2025/07/why-the-commissions-decision-undermines-the-goals-of-the-dma/ (Accessed: 17 November 2026).
- Rijo, L. (2025) ‘Meta announces generative AI advances for advertisers at Cannes Lions’. PPC Land, 21 June 2025. Available at: https://ppc.land/meta-announces-generative-ai-advances-for-advertisers-at-cannes-lions/ (Accessed: 17 November 2026).
- Hutchinson, A. (2026) ‘Meta introduces new AI-powered tools at Advertising Week’. Social Media Today, 6 October 2026. Available at: https://www.socialmediatoday.com/news/meta-introduces-new-ai-powered-tools-at-ad-week-2026/832288/ (Accessed: 17 November 2026).
- Hutchinson, A. (2025) ‘Meta Launches Opportunity Score to All Advertisers’. Social Media Today, 9 June 2025. Available at: https://www.socialmediatoday.com/news/meta-launches-opportunity-score-all-advertisers/750231/ (Accessed: 17 November 2026).
- McCoy, K. (2025) ‘Meta’s AI ad plan raises stakes – even if creative execs are shrugging it off’. Digiday, 9 June 2025. Available at: https://digiday.com/marketing/metas-ai-ad-plan-raises-stakes-even-if-creative-execs-are-shrugging-it-off (Accessed: 17 November 2026).
- Fernandez, A. (2026) ‘Meta’s Advertising Week 2026 ad tools: what is live, in test or coming’. Relevant Audience, 8 October 2026. Available at: https://www.relevantaudience.com/meta/meta-advertising-week-new-york-2026-ad-tools/ (Accessed: 17 November 2026).
- Schiff, A. (2025) ‘For Meta Marketers, Automation Isn’t Always The Advantage (But It’s Complicated)’. AdExchanger, 29 October 2025. Available at: https://www.adexchanger.com/measurement/for-meta-marketers-automation-isnt-always-the-advantage-but-its-complicated/ (Accessed: 17 November 2026).
- Loomer, J. (2025) ‘Meta Andromeda: What It Means for Your Ad Strategy’. Jon Loomer Digital, 19 August 2025 (updated 15 December 2025). Available at: https://www.jonloomer.com/meta-andromeda/ (Accessed: 17 November 2026).
- Vainio, M. (2026) ‘Meta Andromeda and creative strategy: what actually changed and what didn’t’. Precis, 19 July 2026. Available at: https://www.precis.com/resources/meta-andromeda-and-creative-strategy-what-actually-changed-and-what-didnt (Accessed: 17 November 2026).
- Loomer, J. (2025) ‘Meta Andromeda and Creative Diversification: 7 Examples Explained’. Jon Loomer Digital, 20 October 2025 (updated 21 October 2025). Available at: https://www.jonloomer.com/meta-andromeda-creative-diversification/ (Accessed: 17 November 2026).
- Shukla, M. (2026) ‘What Meta’s Andromeda Update Actually Changes – And What It Doesn’t’. AdExchanger, 28 January 2026. Available at: https://www.adexchanger.com/data-driven-thinking/what-metas-andromeda-update-actually-changes-and-what-it-doesnt/ (Accessed: 17 November 2026).
- Loomer, J. (2024) ‘Results: Testing Quality Leads from Instant Forms vs. Website’. Jon Loomer Digital, 6 May 2024. Available at: https://www.jonloomer.com/testing-quality-leads/ (Accessed: 17 November 2026).
- Stuhrenberg, P. (2026) ‘How to Run Facebook Lead Generation Ads Without Low-Quality Leads’. Directive, 19 June 2026. Available at: https://directiveconsulting.com/blog/facebook-lead-generation-ads-quality-guide/ (Accessed: 17 November 2026).
- HubSpot (2026) ‘Use contact and company lifecycle stages’. HubSpot Knowledge Base, 20 October 2026 (last updated). Available at: https://knowledge.hubspot.com/records/use-lifecycle-stages (Accessed: 17 November 2026).
- HubSpot (2026) ‘Set up and manage object pipelines’. HubSpot Knowledge Base, 22 October 2026 (last updated). Available at: https://knowledge.hubspot.com/object-settings/set-up-and-customize-pipelines (Accessed: 17 November 2026).
- Meta (n.d.) ‘Conversions API for CRM integration’. Meta for Developers, n.d. Available at: https://developers.facebook.com/documentation/ads-commerce/conversions-api/conversion-leads-integration (Accessed: 17 November 2026).
- Meta (2024) ‘Version 20.0’. Meta for Developers (Marketing API Changelog), 21 May 2024. Available at: https://developers.facebook.com/documentation/ads-commerce/marketing-api/marketing-api-changelog/version20.0 (Accessed: 17 November 2026).
- Meta (n.d.) ‘Generate Leads with Online Forms’. Meta for Business, n.d. Available at: https://www.facebook.com/business/ads/ad-objectives/lead-generation/lead-ads-with-forms (Accessed: 17 November 2026).
- Meta (n.d.) ‘Conversions API FAQ’. Meta for Developers, n.d. Available at: https://developers.facebook.com/documentation/ads-commerce/conversions-api/conversion-leads-integration/faq.md (Accessed: 17 November 2026).
- Meta (n.d.) ‘Payload Specification’. Meta for Developers, n.d. Available at: https://developers.facebook.com/documentation/ads-commerce/conversions-api/conversion-leads-integration/payload-specification.md (Accessed: 17 November 2026).
- Meta (n.d.) ‘5: Configure Your Sales Funnel’. Meta for Developers, n.d. Available at: https://developers.facebook.com/documentation/ads-commerce/conversions-api/conversion-leads-integration/crm-integration/5-configure-your-sales-funnel (Accessed: 17 November 2026).
- HubSpot (2026) ‘Create and sync ad conversion events with your Meta Ads accounts using Meta’s Conversion API’. HubSpot Knowledge Base, 7 October 2026 (last updated). Available at: https://knowledge.hubspot.com/ads/create-and-sync-ad-conversion-events-with-your-meta-ads-accounts-using-metas-conversion-api (Accessed: 17 November 2026).
- Meta (n.d.) ‘Parameters’ [Conversions API]. Meta for Developers, n.d. Available at: https://developers.facebook.com/documentation/ads-commerce/conversions-api/parameters (Accessed: 17 November 2026).
- Meta (n.d.) ‘Conversions API’. Meta for Developers, n.d. Available at: https://developers.facebook.com/documentation/ads-commerce/conversions-api (Accessed: 17 November 2026).
- Kalungi (2026) ‘Conversion Value Optimization: How to Stop Feeding Meta Junk Leads’. Kalungi, 21 August 2026. Available at: https://www.kalungi.com/blog/meta-ads-conversion-value-optimization (Accessed: 17 November 2026).
- Blochberger, L. (2026) ‘Meta Andromeda, GEM und Advantage+ richtig nutzen’. Blck Alpaca, 25 August 2026 (updated 17 October 2026). Available at: https://blckalpaca.at/de/wissen/social-media/paid-social-performance-marketing/meta-andromeda-gem-advantage-plus (Accessed: 17 November 2026).
- Loomer, J. (2026) ‘You Can Control Lead Quality From Meta Ads’ [Podcast]. The Pubcast with Jon Loomer, 27 January 2026. Available at: https://pubcast.jonloomer.com/you-can-control-lead-quality-from-meta-ads/ (Accessed: 17 November 2026).





