Meta is very good at finding people who are likely to respond to an ad. It is much less good at finding people who meet the financial criteria a Registered Investment Advisor actually cares about, and that undercuts the value of wealth management advertising.
An RIA needs more than someone with an interest in retirement, investing, or financial planning. It needs prospects with enough investable assets to make a new relationship worthwhile. Meta cannot reliably tell you whether a given person has $2 million in investable assets. That is the characteristic that matters most to the firm, and it is the one the platform is least able to target.
The situation has also tightened. Since January 2025, advertisers based in or targeting the United States, Canada, and certain European countries have had to designate financial campaigns under Meta's Financial Products and Services special ad category, which removes several of the targeting tools advertisers previously leaned on. Building an RIA acquisition system on those tools is no longer an option.
So the answer is not to hunt for a hidden high-net-worth targeting setting. It is to build a system in which each layer does what it can and the funnel absorbs the rest. Five layers matter: what native targeting still offers, the conversion signals the firm feeds back into the platform, purchased intent data, creative, and qualification at the booking stage.
This is where sophisticated RIA acquisition differs from selecting an audience in Ads Manager.
Why High-Net-Worth Targeting Is Difficult on Meta
A financial advisor might define a good prospect as someone 55 or older with $1 million or more in investable assets, approaching retirement, living in a particular state, and actively considering a change of advisor.
Meta does not provide a dependable targeting combination representing all of that, and it never did. Geography is available. Depending on the account there may be behavioral signals. But there is no investable assets field in Meta, and that matters more than most advertisers realize.
A person can be 62 without substantial investable assets. Someone can follow retirement planning without having enough to become a target client. A senior executive can have a high income and little liquid wealth. Someone interested in investing can be a $20,000 investor or a multimillion-dollar household.
Demographic characteristics are proxies, not proof. Consider an audience of people aged 55 and over interested in retirement planning. It sounds close to an RIA's ideal prospect, and it contains wildly different financial circumstances: households managing several million dollars, people retiring on modest savings, and people who simply consume retirement content. The same applies to occupation. A physician, attorney, business owner, or executive is more likely than average to hold substantial assets, and occupation is still not a balance sheet.
For an RIA with $500 million to $5 billion in AUM, the objective is not to pretend Meta can identify wealth. It is to give the platform increasingly useful signals while moving the actual financial qualification closer to a point where it can be measured.
The Financial Products and Services Category Changed the Rules
In October 2024, Meta expanded its existing credit ad category to cover financial products and services, a category that includes insurance, bank accounts, investment services, and payment services. In January 2025 it began enforcing that category for advertisers based in or targeting the United States, Canada, and certain European countries.
Any RIA running acquisition campaigns in the United States should assume this applies and should confirm how its own campaigns are being classified inside the ad account, because the practical consequences are significant.
Targeting options restricted under the category include age, gender, ZIP or postal code, exclusion targeting, lookalike audiences, and saved audiences. Certain interests are unavailable for audience creation. Targeting by city or pin-drop location requires an expanded radius.
Read that list against how advisor campaigns were historically built and the problem is obvious. Age brackets, lookalikes from a customer list, and excluding existing clients were the three pillars of conventional RIA targeting, and all three are on it.
Meta is moving the same way more broadly. Over years the platform has leaned increasingly on machine learning and audience signals, and in 2025 the Andromeda ranking and retrieval system made that shift more pronounced. Andromeda is separate from the category restrictions and points the same direction: increasingly it is signals from the system, rather than filters the advertiser sets, that determine who sees an ad.
The important point is not that options were removed. It is that the role of targeting changed. When an advertiser has granular controls, targeting does most of the filtering before an impression occurs. When those controls are restricted, filtering moves to the conversion data the platform receives, the content delivered, and the qualification process after the click.
That is why an RIA's targeting system has to be designed as an acquisition process rather than a list of settings.
Layer One: What Native Targeting Still Offers
Native targeting is still useful. It is just doing a smaller job than it used to.
Geography is the most reliable remaining control and is often the most important one. If an RIA serves clients within a particular region, geography functions as a hard constraint on where the firm is willing to acquire clients. That matters especially for firms operating under state registration requirements or a business model built around a defined market. Note that under the category, city and pin-drop targeting require an expanded radius and ZIP-level targeting is not available, so plan around states and metro regions rather than tight radii.
Age deserves a direct warning. Age brackets are restricted under the category, which means a firm specializing in retirement planning cannot simply exclude 24-year-olds the way it once could. Where age controls are unavailable, life stage has to be communicated through the creative and confirmed at booking rather than enforced in the audience.
Interests should be treated carefully even where available. An interest in investing, retirement, stocks, or financial planning tells you what a person engages with. It says nothing about what that person has. A targeting combination can look highly specific while missing the characteristic that matters most, and some interests are unavailable for audience creation in this category regardless.
One practical step before any of this: confirm how the ad account is classifying the firm's campaigns. The designation determines which controls are actually available, and advertisers are sometimes surprised to find a campaign classified into it, or surprised to find one that was not. A media buyer can answer that in a few minutes, and the answer changes what the rest of the plan can assume.
Native targeting is the starting layer, not the solution. Treat it as establishing the market, not the prospect.
Layer Two: Feed the System Better Conversion Signals
The second layer used to be lookalike audiences. Under the category, lookalikes and saved audiences are restricted, so the mechanism changes even though the underlying logic does not.
The logic is this: rather than telling Meta what a good prospect theoretically looks like, give it examples of people who behaved like good prospects. What changes is that the firm now does this through the conversion events it optimizes toward and reports back, rather than by building an audience object from a customer list.
That makes event quality the whole ballgame. If the campaign optimizes toward page views, the system learns what page viewers have in common, which includes people who clicked accidentally, read an article, looked at a job posting, or visited an employee bio. The signal is diluted. If the campaign optimizes toward booked appointments, the system receives a far more commercially relevant indication of who to find.
Deeper events carry better information. A website visitor has demonstrated interest. A booked appointment has demonstrated intent. The acquisition objective for a high-end RIA is not traffic. It is a qualified conversation.
Deeper events are also rarer. A firm can define events representing form submissions, booked appointments, attended appointments, qualified opportunities, and closed clients. The further down the funnel the event sits, the closer it is to the business outcome and the less volume it produces, which is a real constraint on how quickly the system can learn.
The plumbing has to exist. The CRM, booking system, website, and ad platform cannot operate as disconnected systems. The firm needs a way to identify meaningful downstream actions and send those signals back, typically through server-side conversion reporting. Without that, the platform is optimizing toward whatever shallow event it can see.
Volume and patience both matter. A campaign optimizing toward a deep event needs enough of those events for the system to learn anything, and an RIA generating a handful of appointments a week will take longer to get there than a consumer advertiser generating hundreds of purchases a day. That is an argument for feeding the deepest event the firm can supply at reasonable volume rather than automatically choosing the deepest event that exists.
This is also why data architecture is now a targeting question rather than an IT question. The quality of the signal a firm can send is the main lever it still controls inside the platform.
Layer Three: Purchased Intent Data
The third layer addresses what neither geography nor conversion signals can solve: identifying people whose recent behavior indicates an active financial need.
Take people over 55 with $1 million or more in net worth who searched for a retirement planner or an annuity in the past 10 days. That is different from targeting people with an interest in retirement. The first combines financial traits with recent behavior. The second picks up an interest.
Providers outside Meta can build audiences from data points the platform does not expose, and the value comes partly from recency. Someone who searched for a retirement planner this week is behaving differently from someone who did so six months ago, so providers refresh audiences as new signals arrive and old ones lose relevance. An audience built on stale intent data loses most of its usefulness.
Two cautions. First, confirm what can actually be activated. Custom audience uploads and their availability for financial campaigns depend on the category designation and current platform rules, so verify inside the ad account before building a strategy around this layer rather than assuming the integration will work.
Second, treat intent as a signal rather than a qualification. A provider's definition of net worth may differ from the firm's definition of investable assets, and someone with a $1 million net worth may have very little available for an advisory relationship. Someone searching for an annuity may want information rather than an advisor. Third-party data also contains errors. Use it as an input, then judge it on the quality of the appointments it produces.
This layer earns its keep when the market is small. An RIA is not trying to reach everyone. It is trying to reach a narrow band of households, which means the firm cannot afford to cast an infinitely wide net and wait for the right people to surface. External signals that Meta cannot supply are worth paying for precisely because the addressable pool is limited, and worth dropping quickly if the appointments they produce look no different from everything else.
Layer Four: Creative Does the Targeting the Platform Cannot
As platform targeting narrows, creative takes on more of the job. This is the most overlooked idea in Meta advertising for RIAs, and it matters more now than it did two years ago.
An ad does not simply attract attention. It tells the viewer who the message is for. The subject matter, language, examples, and financial problem described all pre-qualify the audience before anyone clicks.
Imagine an ad that opens by saying retiring with $2 million is a different planning problem from retiring with $200,000, then offers three decisions high-asset households get wrong in the five years before retirement. That message will not appeal equally to everyone. It is naturally more relevant to people with substantial assets approaching retirement, and the creative is performing part of the targeting function. Compare it with are you ready for retirement, schedule a free consultation, which casts a much wider net.
Specificity replaces some of the lost precision. The advertiser may no longer be able to tell Meta which households to find. It can tell prospects exactly what kind of financial problem the firm solves, which changes who chooses to click. The right person sees a problem they recognize. The wrong person sees something that does not feel relevant enough to continue.
For RIAs this works particularly well, because the firm's expertise is usually expressible as a problem: managing concentrated equity positions, planning around a business sale, coordinating retirement income across multiple accounts, evaluating a major liquidity event, or coordinating tax and investment decisions for high-asset households.
Creative also has to keep moving. Self-selection depends on the message staying sharp, and a concept that filters well in month one filters less well once the audience has seen it repeatedly. Treat creative as a pipeline rather than a launch, with new problems entering as older ones fatigue.
The point is not to insert a wealth claim into every ad. It is to make the problem specific enough that the right prospect recognizes it.
Layer Five: Qualification at Booking
Every targeting system eventually reaches the same limit. Meta cannot verify someone's investable assets, and under the current category it cannot even filter on age. So qualification cannot end when someone clicks.
The booking process performs the final filtering. A firm can ask qualifying questions before allowing someone to schedule, including investable assets expressed as ranges that reflect its minimum household size. The exact questions and ranges should follow the firm's business model and compliance requirements.
This creates the distinction that organizes everything else. Targeting tries to put the ad in front of the right people. Qualification determines whether an individual prospect meets the firm's requirements. Those are different jobs, and the second one has become more important as the first has become less precise.
Do not strip qualification to protect booking volume. There is constant pressure to shorten the form, because every question costs some percentage of bookings. For an RIA that trade is usually backwards. A calendar full of people who cannot meet the firm's minimum burns the most expensive hours in the business while making the marketing dashboard look healthy. The question to apply to each field is whether the answer changes what happens next, in qualification, routing, or preparation. If it does not, it does not belong.
If the platform cannot determine whether someone has $2 million in investable assets, the firm does not need to force it to. It can ask the prospect.
That leads to a strategic principle worth stating plainly: not every impression needs to come from a perfectly qualified person. What matters is that the complete system separates potential prospects from people clearly outside the firm's market. Targeting identifies likely relevant people. Creative makes the problem specific and encourages self-selection. The landing page sets expectations. The booking form asks the qualifying questions. The appointment lets the advisor determine genuine fit. Each stage removes a different kind of uncertainty, which is far more realistic than expecting one audience setting to identify the firm's ideal household.
Geography Is a Strategic Decision
Geography deserves separate attention, partly because it is now one of the few controls that survived and partly because the right approach depends on the firm's business model. A local RIA, a multi-state RIA, and a nationwide RIA should not use the same structure.
Radius targeting makes sense when the target market is concentrated around an office or a specific local market, and it can keep acquisition focused where the firm has an established presence, referral network, or in-person service model. Remember that the category requires an expanded radius for city and pin-drop targeting, so very tight local targeting is not available the way it once was. The radius should follow the actual service model rather than an instinct to make the audience smaller. If clients routinely travel an hour to meet the advisor, a narrow radius excludes perfectly viable households.
State-level targeting is appropriate when registration or service constraints limit where the firm can actively solicit. In that situation geography is not a marketing preference. It is part of the operating and compliance framework, and the setup should reflect the states the firm is actually prepared and permitted to serve. It is also worth documenting why the campaign was configured the way it was, because the person who set it up is not always the person who has to explain it later.
Nationwide targeting makes sense for a genuinely national service model. A firm serving clients virtually across the country gains nothing from artificially restricting itself to a few markets. The distinction that matters is between a firm that could technically serve someone anywhere and a firm whose strategy is built around serving people anywhere. Geography should follow the second.
Suppression Has to Move Out of the Ad Platform
Targeting is not only about who should see an ad. It is also about who should not, and this is where the category restrictions bite hardest.
Exclusion targeting is restricted, which means the familiar approach of uploading a client list and excluding it from acquisition campaigns is not available in the way firms are used to. The underlying need has not gone away. Existing clients should not consume budget intended to generate new relationships. People who have already converted should not be treated as fresh prospects, because continuing to do so distorts campaign economics and wastes impressions.
Since the platform cannot do this filtering, the firm has to absorb it elsewhere. Suppression moves into the funnel and the operation: booking forms that identify existing relationships before an appointment is confirmed, CRM routing that catches current clients who book anyway, creative and offers aimed at problems prospective clients have rather than ones existing clients are already being served on, and campaign reporting that reconciles new appointments against the existing household list rather than assuming every booking is net new.
Retargeting still needs its own structure. Someone who reached a landing page and did not book is in a different position from someone who has never encountered the firm, and the message that serves one does not serve the other. Keep those efforts separate from cold acquisition rather than letting them compete for the same budget and blur the same reporting.
It is less elegant than an audience exclusion and does the same work. The principle holds either way: do not pay repeatedly to acquire the same person, and do not count a current client as an acquisition.
Putting the Layers Together, and What to Measure
The layers are not competing strategies. Geography establishes the market. Conversion signals tell the system what a valuable prospect actually did. Purchased intent adds outside information where it can be activated. Creative produces self-selection. Booking qualification asks the questions the platform cannot answer. The platform does not need to know everything. It needs enough useful information to find people likely to respond, while the rest of the system handles qualification.
Measuring that system is where most RIAs go wrong, because the tempting metrics sit at the top. Click-through rate is easy to see. Cost per click is easy to compare. Lead volume is easy to report. None of them tells a principal whether the targeting is producing the households the firm wants.
The comparison that matters happens further down. When testing geography against purchased intent, or one conversion event against another, evaluate the appointments each produces. Are prospects meeting the minimum asset criteria? Are they attending? Are advisors calling the conversations qualified? Do certain audiences consistently produce better downstream opportunities? Firms differ in market, data quality, and value proposition, so the job is to find which signals produce useful prospects for this RIA rather than assume one method always wins.
There is a natural desire to find the perfect high-net-worth audience. On Meta that is the wrong objective, and the category restrictions have made it an even less dependable foundation than it was. The better approach is to build around the information you actually have: use native targeting where it still works, send better conversion signals, add purchased intent where it can genuinely be activated, make the creative specific enough that the right prospects self-select, and ask the qualifying questions at booking that the platform cannot answer.
High-net-worth acquisition is not won by finding a magical audience setting. It is won by combining imperfect signals until the funnel produces a strong enough qualification process. And as Meta continues moving toward broader, machine-driven audience selection, the firms with the best first-party signals, clearest creative, and strongest qualification will have more useful information to give the platform than firms still relying on demographic proxies for wealth.
