The NEW BEST Way to Analyze Facebook Ads in 2026 (4Pi after Andromeda) — Transcript

Full transcript

1. 0:00Tired of pouring money into Meta ads and
2. 0:01feeling like you're getting nowhere?
3. 0:02You're not alone, and you're not stuck.
4. 0:05I get it. It's frustrating, but here's
5. 0:07the thing. Meta's algorithm isn't
6. 0:10actually a black box trying to take your
7. 0:11money, and once you know what to look
8. 0:13for, [music]
9. 0:14everything changes. In this video, I'm
10. 0:16going to show you the four simple
11. 0:18metrics that will let you understand
12. 0:20exactly what it's doing, so you can quit
13. 0:22worrying about why things aren't working
14. 0:24and start seeing real, steady growth.
15. 0:31Spend isn't just a budget setting. It's
16. 0:34the clearest way to see what decisions
17. 0:36Meta is making about your ad. Meta sells
18. 0:39attention, and attention is precious and
19. 0:42fragile. Before your ad earns real
20. 0:43spend, it has to pass through two gates.
21. 0:45Lattice balances these are experience
22. 0:47with advertiser objectives. Basically,
23. 0:50should this person see an ad at all? And
24. 0:52Andromeda is the filter, choosing which
25. 0:54of the millions of ads competing for
26. 0:56that impression deserve a real chance to
27. 0:59earn distribution. So, when your ad
28. 1:01starts earning spend, it's usually
29. 1:03because those systems are seeing signals
30. 1:05like people stopping their scroll. They
31. 1:07watch. They click. They convert. Meta is
32. 1:10saying, "Users like this experience.
33. 1:13We're going to show it to more people."
34. 1:15And that is why ads getting the most
35. 1:17spend can sometimes have a slightly
36. 1:20worse CPA. Because when Meta is spending
37. 1:22higher, it's pushing your ad past the
38. 1:23easy conversions and into new, colder
39. 1:25people that fill the funnel. Spend is a
40. 1:28measure of preference, ranking ads by
41. 1:30how well they meet both your business
42. 1:32objectives and Meta's, which brings us
43. 1:34to the most misunderstood metric in Meta
44. 1:37ads.
45. 1:42It's easy to see frequency rising and
46. 1:43panic and call it fatigue, but frequency
47. 1:46doesn't really show fatigue. Frequency
48. 1:48is telling you how Meta is using an ad
49. 1:50in the funnel. Here's a simple way I
50. 1:52want you to think about it. When you're
51. 1:54looking at daily frequency, when it's at
52. 1:56a 1.05, that means 5% of people saw that
53. 2:00ad twice. When your daily frequency is a
54. 2:021.95,
55. 2:04that means about 95% of people have seen
56. 2:07that ad twice. Lower frequency usually
57. 2:10means the ad is being used to find new
58. 2:13people. Higher frequency usually means
59. 2:15Meta is using that to retarget
60. 2:17individuals. There are three common
61. 2:19reasons your frequency will climb.
62. 2:22Number one is that cold audiences don't
63. 2:24like the ad. The engagement and
64. 2:26conversion signals aren't strong enough.
65. 2:28So, Facebook just stops expanding to new
66. 2:30people. Number two, you keep resetting
67. 2:33the machine. If you're launching too
68. 2:35many ads, changing budgets, bids, or
69. 2:37audiences, so nothing gets enough stable
70. 2:39data. Meta can't confidently find new
71. 2:42people, so it leans on the people it
72. 2:44already understands. Those two usually
73. 2:46show up as high frequency and low spend
74. 2:48because the system is protecting the
75. 2:50user experience. A third option is an ad
76. 2:53that converts really well on warm
77. 2:54audiences. In that case, it can still
78. 2:57earn spend even with the high frequency
79. 2:59because it's doing a great job later in
80. 3:02the journey. Remember that creative does
81. 3:04the targeting. Frequency is how you can
82. 3:07see who's being targeted. So, now that
83. 3:09we know what the algorithm wants to do
84. 3:11with every impression we buy, we have to
85. 3:13know how much those impressions cost and
86. 3:15how good they are. Now, brings us to
87. 3:18CPM. [music]
88. 3:22Your ads aren't only competing against
89. 3:24other ads, they're competing against
90. 3:26organic content, too. Your friends,
91. 3:28creators, memes, whatever. In Meta's
92. 3:30business model, where attention is the
93. 3:32product, it's in their best interest to
94. 3:35show people content they'll engage with
95. 3:37and ads for products they want from
96. 3:39brands they can actually trust. There's
97. 3:41a second factor, too, supply and demand.
98. 3:44The audiences closer to buying are often
99. 3:46the same audiences your competitors
100. 3:47want. So, there's less supply and more
101. 3:49demand, that equals higher CPM. But,
102. 3:52Meta doesn't just sell impressions to
103. 3:54the highest bidder. It's optimizing for
104. 3:56the outcome and the experience. Ads with
105. 3:59higher quality experiences can earn
106. 4:02lower CPMs. And as users get closer to
107. 4:05buying, Meta gets more selective about
108. 4:07what ads it will show them. It cares
109. 4:10about what it can predict will happen
110. 4:13after the impression. If your ads create
111. 4:16a great user experience, you can often
112. 4:18be charged less to reach people more
113. 4:21likely to convert, even in competitive
114. 4:23audiences. And if your ads create a bad
115. 4:25experience, you often have to pay more
116. 4:28just to get delivery. Yes, CPM is the
117. 4:31cost of attention, but more importantly,
118. 4:33it's a read on the quality and
119. 4:35competitiveness
120. 4:37of that attention.
121. 4:42That brings us to the other really
122. 4:44important cost metric, cost per result.
123. 4:46A lot of people use it to decide whether
124. 4:48an ad is good or bad, and honestly,
125. 4:49whether they're actually good or bad at
126. 4:52the job of running ads. Lower funnel ads
127. 4:54can look incredible. Meanwhile, upper
128. 4:56funnel ads almost always look worse on a
129. 4:58CPA basis because they're doing heavier
130. 5:01work earlier. They grab attention, they
131. 5:03build intent, and earn trust, and prove
132. 5:05authority. A lot of teams optimize by
133. 5:08chasing a better CPA and a higher ROAS.
134. 5:10What they're really doing is trying to
135. 5:11spend as much of the budget as possible
136. 5:13at the bottom of the funnel, not by
137. 5:15audience, but by how Meta is using their
138. 5:18ads, which is a very important
139. 5:21distinction. They effectively cut
140. 5:23prospecting because it looks expensive.
141. 5:25They keep only the ads that close. And
142. 5:27so, your CPA drops, and your ROAS goes
143. 5:29up, and everybody celebrates.
144. 5:31But, then the funnel starts shrinking.
145. 5:33Your frequency climbs, your CPMs climb,
146. 5:36your CAC climbs, and eventually revenue
147. 5:38stalls because you've optimized for
148. 5:40efficiency instead of growth. Cost per
149. 5:42result is not a measure of how good an
150. 5:44ad is. It's a measure of how much that
151. 5:46ad contributes to the conversion
152. 5:48actually happening. The lower the cost,
153. 5:50the less work it probably did because it
154. 5:52showed up at the end. And the higher the
155. 5:55cost, the more work it might be doing
156. 5:57because it's creating demand earlier in
157. 6:00the journey. So now that we've covered
158. 6:02all the metrics, let's talk about how to
159. 6:04put them to work.
160. 6:09When you use these four metrics, spend,
161. 6:11frequency, CPM, and cost per result to
162. 6:13understand how the machine is using your
163. 6:15ads, you are conducting a 4PI analysis.
164. 6:18Let me give you three of the most common
165. 6:20patterns that I look for. Two examples
166. 6:23of an ad that is doing its job well, and
167. 6:25one huge red flag that tells you what to
168. 6:27pause or replace with creative testing.
169. 6:31Example number one, a great upper funnel
170. 6:34ad prospecting. The spend is above
171. 6:36average. It's spending more than the
172. 6:37other ads. The frequency is below
173. 6:39average. Your CPM is below average
174. 6:41compared to your other ads. It's a good
175. 6:43attention experience. And the cost per
176. 6:45result, it's probably above average, and
177. 6:47that's normal. Cost per result can be
178. 6:49higher as long as it's not so high that
179. 6:52the ad becomes a liability relative to
180. 6:54the spend that it's taking. This is the
181. 6:56ad that creates a great first touch, and
182. 6:58it should set up the ads that come later
183. 7:01in the funnel. Example number two is a
184. 7:03great lower funnel ad, retargeting and
185. 7:05closing. A frequency that's higher than
186. 7:07average, a CPM that's higher than
187. 7:08average, and a cost per result that's
188. 7:10better than average. Spend should be
189. 7:13meaningful enough that it actually
190. 7:15contributes. We don't want it to spend
191. 7:17just enough to get one or two sales a
192. 7:18week. It should be the highest spending
193. 7:20thing that has a greater than average
194. 7:22CPM and frequency. When these ads are a
195. 7:24logical second touch to the best
196. 7:26prospecting ad, you'll often see them
197. 7:28get even more efficient as the funnel
198. 7:30fills. Let's talk about the red flag.
199. 7:32Expensive results, and the machine still
200. 7:35loves it. If an ad has a bad cost per
201. 7:37result, but it's still earning a ton of
202. 7:39spend, it might be amazing at attracting
203. 7:42the wrong kind of attention. It's
204. 7:44getting distribution, but it's pulling
205. 7:46in people with no intent to buy. So, it
206. 7:48becomes a liability to the bottom line.
207. 7:51And another version of this same red
208. 7:53flag is frequency above average, CPM
209. 7:55above average, but the cost per result
210. 7:57is no better than the prospecting ads.
211. 7:59That usually means the ad appeals to a
212. 8:01warm audience, but doesn't convert them.
213. 8:04So, it soaks up budget that could be
214. 8:06going to an ad that actually improves
215. 8:09efficiency and drives sales. Now that
216. 8:11you can spot these patterns, the next
217. 8:13step is simple. What do you turn off?
218. 8:15What do you keep? And what do you test
219. 8:17next based on your 4 PI? And that is
220. 8:20what we're doing in the action items.
221. 8:22Now, I hope you're enjoying watching
222. 8:23this video as much as I enjoyed making
223. 8:25it. Please go ahead and like and
224. 8:26subscribe so you don't miss anything
225. 8:27else. If you want to learn more about
226. 8:29the 4 PI analysis, if you want to dive
227. 8:31all the way into the deep end, we just
228. 8:33dropped a brand new course in Disruptor
229. 8:35Academy. There's a link right there so
230. 8:36you can join. All right, with that being
231. 8:38said, let's get back to it.
232. 8:44Data is only ever as valuable as it is
233. 8:46actionable. So, what actions does a 4 PI
234. 8:49analysis tell us to do? There are really
235. 8:51only two logical outcomes if you want to
236. 8:53make the account better. Outcome number
237. 8:55one is optimization by subtraction. This
238. 8:57is the simplest one and honestly, it's
239. 8:59often the fastest way to scale results.
240. 9:01Your goal is to remove an ad that's
241. 9:03hurting the system. So, the budget
242. 9:05naturally flows to the ads that are
243. 9:07actually doing the job. You're scaling
244. 9:08the winners without touching the budget.
245. 9:10Because remember, we're not trying to
246. 9:11find one winning ad to isolate, scale,
247. 9:14and abuse until it breaks. We're trying
248. 9:16to build the best team of ads that
249. 9:19produces a repeatable, stable, and
250. 9:21scalable outcome. So, if you're running
251. 9:23a lot of ads, maybe more than eight,
252. 9:26start here. Find the ads that are taking
253. 9:28meaningful spend and have clear red
254. 9:31flags and start turning those off.
255. 9:33Eliminating the liability is the easiest
256. 9:36way to get better. And outcome number
257. 9:38two, improve the team by filling the
258. 9:41gaps. If you don't have a clear red
259. 9:42flag, you don't have a pause stuff
260. 9:45problem. You have a coverage problem in
261. 9:48the funnel. Here are the most common
262. 9:49ways it shows up. First, everything
263. 9:52looks fine until you raise budget. And
264. 9:54when you scale, CPM, frequency, and cost
265. 9:57per results all get worse across the
266. 10:00board. That usually means you're not
267. 10:02filling the funnel effectively enough.
268. 10:04So, your creative testing needs to focus
269. 10:06on stronger upper funnel ads. Number
270. 10:09two, when you scale, one prospecting ad
271. 10:11soaks up all the spend and CPR gets
272. 10:13worse and worse. And that's usually an
273. 10:15ad people love to watch, but it's
274. 10:17attracting the wrong kind of attention
275. 10:19at scale. So, the fix isn't more budget
276. 10:21tweaks and bidding modifications. The
277. 10:23fix is to replace or iterate on that
278. 10:26prospecting ad to give the machine
279. 10:28better choices. And number three,
280. 10:29nothing behaves like a lower funnel ad
281. 10:31that can earn spend. So, you don't have
282. 10:33a closer. So, you need to test for an ad
283. 10:35with a higher frequency, higher CPM, and
284. 10:36a better CPR. Something that can convert
285. 10:39the warm traffic that you're buying more
286. 10:41efficiently. The easiest way to do that
287. 10:43is to take a strong prospecting idea and
288. 10:45make a second touch version. Same core
289. 10:47message, but built for someone who
290. 10:49already saw the first ad.
291. 10:56Your job isn't to find one winning ad
292. 10:59and abuse it until it breaks. Ad fatigue
293. 11:02is almost always operator error. Your
294. 11:05job is to get ads that are working
295. 11:07together so the system stays balanced,
296. 11:09profitable, and scalable. Because that
297. 11:11balance is what lets you answer the only
298. 11:14question that matters. Can I spend more
299. 11:16money with a yes? A 4 PI analysis turns
300. 11:19your ads manager into a decision engine.
301. 11:22So, now you know how to look at all the
302. 11:23ads. I got to show you how to test more
303. 11:25effectively so you can cover those gaps
304. 11:27and scale your account. And that's
305. 11:28exactly what this next video does.
306. 11:33>> [music]
307. 11:40[music]