At a $2,500/month ad budget — Nova's recommended minimum to generate meaningful data — the money works best concentrated, not scattered. A workable starting split is roughly $1,500–$1,800 (60–70%) into Google Search and Local Service Ads (LSA) targeting high-intent, near-application-ready searches; $500–$700 (20–25%) into Meta retargeting aimed at past clients and referral-partner audiences, structured around the credit special ad category; and $200–$300 (10%) held as a testing reserve to shift toward whichever channel is producing lower cost-per-application. Management fees sit outside this $2,500 and scale with total spend — the full $2,500 goes to media.
Why Shouldn't the First $2,500 Be Split Across Five Channels?
Because each channel needs a minimum volume of clicks and conversions before the algorithm — or you — can tell what's working. Split $2,500 evenly across Google, Meta, TikTok, LinkedIn, and Spotify and every channel lands at $500/month, which is rarely enough for Google's auction to exit learning mode or for Meta's delivery system to find a reliable pattern in a specialized, compliance-restricted audience. Two channels running at real volume beat five channels running on fumes.
Where Does the Google Ads Portion of the Budget Go?
Most of it goes to a small set of branded and high-intent, application-stage keywords rather than broad awareness terms. For a loan officer, that typically means: your name plus "mortgage broker," "loan officer," or "pre-approval"; local terms like "[city] mortgage broker" or "[city] home loan pre-approval"; and buyer-stage phrases like "mortgage pre-approval online" or "conventional loan pre-approval." LSA budget (Google's pay-per-lead format, verified with an NMLS-checked profile) is worth testing alongside Search because it puts a phone number and reviews directly in front of the searcher with less setup than a full Search campaign — and at $2,500/month, LSA can absorb a meaningful chunk of the Google allocation on its own before Search spend is added.
What Does the Meta Portion of the Budget Buy Under the Special Ad Category?
It buys retargeting and warm-audience reach, not broad prospecting. Meta classifies mortgage and credit advertising under its credit special ad category, which strips out age, gender, and detailed demographic targeting and narrows geographic targeting to a minimum radius. That makes cold lookalike audiences weaker for this niche than they are for, say, a roofing company. What still works well within those restrictions: retargeting people who've already visited your site or landing page, and custom audiences built from uploaded lists — past clients, referral partners' contact lists (with proper consent), or webinar/event attendees. That's where the Meta dollars in a $2,500 budget are best spent, rather than on cold reach.
What About Landing Pages, CRM, and Tracking — Does That Come Out of the $2,500?
No, and this is where budgets quietly leak. A TCPA-compliant lead form, visible NMLS ID, clear rate-claim disclosures, and call tracking (something like CallRail) are infrastructure, not media — they should be in place before the $2,500 starts spending, not funded out of it. Skip this step and the clicks still cost the same, but a meaningful share of the leads they generate become unusable or noncompliant, which is a worse outcome than a smaller, cleaner budget. This is a marketing-structure decision, not legal advice — a loan officer's compliance counsel should sign off on the actual consent language and disclosures used on the landing page and forms.
Two Ways to Split $2,500
| Model | Google (Search + LSA) | Meta (retargeting/custom audiences) | Best for |
|---|---|---|---|
| LSA-heavy | $2,000 (mostly LSA) | $500 | Loan officers who want phone calls fast and have limited time to manage a Search campaign |
| Search + Meta split | $1,700 (Search-weighted) | $800 | Loan officers with an existing database and referral network to retarget against |
When Is $2,500 Not Enough to Run Two Channels?
When the loan officer's market is a high-cost metro where per-click prices push the realistic minimum for even one channel above $2,000. In those markets, it's usually better to run one channel — typically LSA or a tightly scoped Search campaign — at full strength for 60–90 days, build a data set, and add a second channel once the first is producing consistent cost-per-application numbers. Adding a second underfunded channel too early just recreates the five-way split problem with two channels instead of five.
Frequently asked questions
Should a loan officer start with Google or Meta?
Google, in most cases — Search and LSA capture people actively looking for a broker or pre-approval, which converts faster than Meta's audience-building approach. Meta becomes more valuable once there's a database or referral list to retarget, which is why it's usually the smaller share of the first $2,500 rather than the starting point.
Does the Meta special ad category make Facebook ads not worth it for mortgage?
It changes what Meta is good for — it narrows broad prospecting but doesn't block retargeting or custom-audience use, which are where most of the value sits anyway for a loan officer's budget. The practical effect is less about whether to use Meta and more about which audiences to fund with it.
How much of the $2,500 should go to testing new keywords or audiences?
A reserve of roughly 10% ($200–$300) is enough to test a new keyword theme or a new custom audience without disrupting the core campaigns. Anything larger than that at a $2,500 monthly budget usually pulls too much volume away from the channels that are already producing.
Does Nova have mortgage case-study numbers to back this split?
Not yet — Nova doesn't have a published mortgage case study, so this allocation is offered as a starting framework based on how Google's and Meta's platforms behave in this compliance-restricted niche, not as a proven result. A free strategy call with Nova Marketing (novamarketing.ai) is the best way to pressure-test the split against a specific market and referral network.