All HockeyStack Labs reports are done using anonymized HockeyStack customer data. We did not partner with anyone on the creation of this report and it was not sponsored by a vendor.
During my time at Cognism, we were on the brink of killing Facebook back in late 2022. The numbers were definitely not looking good, and the consensus was to reallocate that budget to another channel. It seemed like the easy way out. It would help us focus on fewer channels and improve team capacity and reporting. But this would also mean we’d only rely on Google and Linkedin—back then we hadn’t started experimenting with Youtube, Spotify, Bing, and Reddit yet. Youtube was going to be our fourth channel after: Linkedin, Google, and Facebook, and we were set to start experimenting with it in the next few weeks.
I didn’t want to kill Facebook right away—maybe it was a gut feeling, but I still had hope. I said, one last dance. Since everyone else wanted to kill Facebook, I had no pressure. It couldn’t get any worse, so I changed the entire account structure in a day: no more mirroring the Linkedin setup, fewer campaigns, completely different ads, but most importantly, different objectives.
The pipeline impact didn’t happen quickly, but we saw an incredible uptick in website visitors and high-intent page visits in a week of two. This was my indicator to keep it running longer. After a quarter, the Facebook pipeline increased by 500% with only a 5% budget increase. The first two quarters of 2023 were a golden rush for B2B ads on Facebook. In the third quarter, numbers started to flatten. By the fourth quarter, they got really low. We were still generating MQLs, but who cares about MQLs if they don’t convert?
We gradually decreased the budget experiment by experiment, as most of them were failing. By mid-Q4 2023, Youtube and Bing were getting more budget than Facebook; by late Q4, Facebook only had retargeting campaigns. When I wrote the 2023 Recap report for HockeyStack, Facebook numbers looked promising and these numbers were pretty similar to what we had at Cognism. Then we did the Q1 Recap 2024 report, and I saw that the budget allocation for Facebook was getting even bigger than the 2023 average. But by the Q2 Recap 2024 report, we saw that the budget for Facebook had plummeted, and the numbers were simply devastating. The second quarter of 2024 marked the downfall of Facebook; after seeing those numbers, I wanted to dive deeper and analyze what had been happening with Facebook.
Methodology
MQL: Direct form submissions such as demo, pricing page, and contact us. This doesn’t include any ebook downloads, webinar registrations or anything like that. Currently there’s an inflation of terms in B2B marketing, some say ‘hand raisers’, some say ‘high intent conversions’. I’m calling this MQL simply because I like the sound of it.
SQL: Pipeline created. Every company has different definitions, but we unified this on the backend and used the SQL definition for when the pipeline is actually created.
Sales Cycle: Starts from the date of deal created, not from MQL created. The reason for this is that since we included both inbound and outbound, and outbound doesn’t start from the MQL level, it would have skewed the average.
Deal Size: Total new deal amount added divided by the number of qualified deals (also known as Deal Value)
ACV: Total new business revenue added divided by the number of closed-won deals (also known as Average Contract Value, Selling Price)
CW: Closed won deals, net new business, new revenue.
Sample Size: 41 B2B SaaS companies & over $18M ad spend
Attribution Model: Position-based because it offers a balanced view on the importance of each touchpoint in a customer's journey towards a conversion. This method assigns more credit to the first and last interactions and distributes the remaining 20% among the middle interactions.
Sample Description:
- At least $25k monthly paid spend per company
-At least 1% budget allocation per company to Facebook
- From $6M ARR to $2B ARR; average ACV from $8.5K to $150K.
- 75% NAM, 20% UK, 5% Europe
Part I: Budget Allocation
In the Q2 2024 recap report, we found that the budget allocation for Facebook dropped from 16% to 3.04% from Q1 to Q2, while the 2023 average was close to 5%. To understand the trends better, we decided to start from the last quarter of 2022—the quarter I began seeing results on Facebook while at Cognism—looking at seven quarters from the start of Q4-2022 to the end of Q2-2024.
From this t=0 point, we clearly see a positive trend in Facebook spending, with budget allocation increasing quarter over quarter and peaking in Q1 2024. However, in Q2 2024, it reached its lowest position in the last seven quarters. So, the first thing we need to consider is, could it be a coincidence? Can we assume this will change this quarter? To understand this, we need to first look at what happened in Q1-24, what the numbers show, and what could have caused this shift across all of our customers. We’ll examine this in the next part—right now, I’m just trying to understand the change in in-platform metrics and spend.
While we’re seeing the lowest budget allocation in the last two years, I see another clue—the CPC in Q1 2024 was the highest in the last two years; but that’s not all. The CPC in Q2 2024 was almost twice as high as the Q1 average. Put another way, the CPC of Facebook in Q1 was 2.3x higher than the average of the last two years, and the CPC in Q2 was 1.9x higher than in Q1. This data clearly shows us how expensive Facebook got in 2024. Obviously, being expensive is not necessarily a massive issue—Linkedin is also expensive—but if it’s costly, it needs to be delivering results, right? Well… But before that, one last thing I want to check is the CTR. Has engagement dropped? Do we see any pattern, or does it look like chaos? Can we see anything at all here?
Looking at the CTR level, it doesn’t look as concerning as the CPC part, but that doesn’t mean it looks good. While the average CTR in Q4 2022 and Q1 2023 was 0.52%, we see a huge jump to 1.05% in the rest of 2023; the average CTR in 2023 was 0.92%. In 2024, however, this CTR dropped to 0.86% in the first quarter and then to 0.78% in the second quarter. The average CTR in 2024 is still better than in the last quarter of 2022; however, it’s more than 20% lower than the average CTR in 2023.
Another interesting thing here is the situation with impressions. While everything else is going terribly wrong, we’re seeing more and more impressions. The average number of impressions in 2024 is 2.6x more than the average of 2023. Even though we had lots of impressions in Q1 2024, aligned with the massive budget allocation increase, these impressions didn’t decrease much in Q2 2024 (although the budget allocation decreased massively)—indicating that CPM got significantly lower compared to previous periods. I think this is another clue. These ads are getting impressions, but ROI is getting lower—oops, spoiler—maybe we’re seeing a shift in the Facebook approach. Do these companies all have a different approach now? It’s rather coincidental; if the dataset included 10 companies, maybe a shift in one or two companies would impact the average. But this dataset is large enough that there’s no way a shift in one or two companies would have a massive impact. I don’t think all these companies changed their approach at the same time, causing a huge peak in the number of impressions. But the data is here—while all other numbers are going down, somehow these Facebook ads are being delivered to more people at a cheaper cost.
Part II: The MQL:SQL Conversion Rate
The MQL:SQL conversion rate highlighted a significant issue with Facebook in the Q2 Recap report. The average MQL:SQL conversion rate on Facebook was 23% in 2023, but it dropped to 13.93% in Q1 2024 and fell further to 12.81% in Q2. Examining this data by quarters reveals the problem more clearly. The average MQL:SQL conversion rate on Facebook was 21.5% in Q4 2022. Throughout 2023, it ranged from 21% to 26.3%. In Q1 2024, it halved and continued to decline in Q2. But, why the decrease? Is it because fewer MQLs are being generated and are less likely to convert, or is it because the increased number of MQLs are less qualified than before?
The answer appears to be a bit of both. The first half of 2023 had a similar number of MQLs as the 2022 average, but in the second half of 2023, companies in our dataset saw a massive increase in MQLs, with numbers rising 4.83x compared to the previous period. This increase continued from H2-2023 into Q1 2024, where the average number of MQLs per company improved by 1.97x compared to the previous period. This meant cheaper MQLs throughout 2023, but not in 2024. As mentioned earlier, the budget allocation for Facebook surged by over 5x in Q1 2024 compared to the 2023 average. However, the number of MQLs increased by only 1.97x, leading to a significant rise in cost per MQL—more than 2x during this period. Furthermore, the drop in conversion rate clearly indicates that while the cost per MQL rose, the pipeline didn’t grow proportionally, meaning the cost per SQL increased and pipeline ROI decreased.
Another crucial point is the change in deal sizes. In Q1 2024, the average deal size from Facebook decreased by 24% compared to the 2023 average, and in Q2 2024, it decreased by another 9%. This means that the average deal size from Facebook dropped by about 30% in the first half of 2024 compared to the 2023 average. Therefore, we’re not only seeing issues with conversion rates and the pipeline but also with deal sizes.
So, why is this happening? I think it’s a combination of economics of scale and perhaps greed. Facebook has always been a supplementary channel; one can't expect it to perform like Linkedin as Facebook is fundamentally different. However, once marketers saw that Facebook was working, they may have decided or may have been pressured to allocate more budget to it. This overinvestment likely drained their audience. Unlike Linkedin, where audiences are regularly updated as people frequently update their profiles, Facebook audiences don’t change as often since users rarely update their job information there. Consequently, once the target market is drained and you hit a certain amount of frequency, it’s hard to replenish. In my opinion, Facebook likely started showing ads to similar audiences, look-alikes. While these look-alikes did convert into MQLs, they were not necessarily ICPs, leading to lower conversion rates from MQLs to SQL. The lack of audience replenishment meant that as the budget allocation increased, the available active audience on Facebook could not sustain it. Metrics deteriorated, causing marketers to lose trust in Facebook, leading to a substantial budget shift back to levels lower than two years prior.
Part III: Revenue
Before coming to a conclusion, I want to examine the revenue trends—because if we see positive changes in close rates and revenue, it could render everything mentioned above redundant - which is very unlikely but we have to be sure. Since we’ve already discussed average deal sizes in the last part, I’ll start by looking at the average contract value. Note that I use ‘deal size’ for all qualified deals and ‘contract value’ specifically for closed-won deals. We’ve observed a decrease in deal size, but what about discount rates?
To recap, the average deal size dropped by about 30% in 2024 compared to the 2023 average. The Q4 2022 average was nearly identical to the 2023 average, with a difference of less than 1.2%. In this context, looking at the average contract value for 2024, we see that Q1 2024 had a 27% lower ACV, meaning there was an additional 3% discount compared to the 2023 average.
This discount rate increased further in Q2 2024. During this period, the average deal size decreased by another 9%, while the average contract value fell by 17%, indicating that salespeople offered 8% more discount to close deals. However, this might not be entirely related to Facebook, as the latest recap report revealed that the average contract value in Q2 2024 was 15.4% lower than in Q1. Thus, although Facebook’s average contract value decreased by 17%, this drop is only slightly different than the overall average decrease.
When it comes to close rates, however, we see minimal change—this might be the only consistently stable metric over the last seven quarters. The difference between Q4 2022 and the 2023 average close rates is less than 5%. In Q1 2024, the close rate decreased by 3% to 12.25%, but it increased to 13.51% in Q2 2024. When we look at the entire data set, the average close rate for Facebook across the entire period is 13.2%.
Conclusion
The data paints a clear picture: Facebook's effectiveness for B2B has significantly declined in 2024. We've seen budget allocations spike and then plummet, CPCs double, and MQL:SQL conversion rates halve. The massive increase in impressions coupled with lower conversion rates and smaller deal sizes suggests we're reaching a broader but less targeted or maxed out audience. This trend likely stems from audience saturation and Facebook's inability to replenish high-quality B2B prospects at the same rate as budget increases. In a nutshell, this analysis reveals one thing: it’s the downfall of Facebook.