Analysing Tech Layoffs
Which function in tech was hit the hardest.
At the time of writing, layoffs recorded by layoffs.fyi had just surpassed the half-million mark across 2581 companies. This astonishing figure shows just how ruthless the tech industry can be. Having been impacted by layoffs myself in Q4 2022, I know firsthand how stressful this time can be. If you're one of the half a million affected, I hope you find a new role soon, if you haven’t already. I've seen heartbreaking posts on LinkedIn from people still looking for employment months after being made redundant.
Heartbreaking as it is, to do this analysis properly, I need to remain emotionally detached to avoid bias.
A few days ago, I saw a LinkedIn post that got me thinking: "Which function in tech has been hit hardest by layoffs?" Was it evenly spread, or were some roles more at risk than others?
So I thought I’d set about answering this question as objectively as I could.
To answer this, I analysed data from layoffs.fyi, which provides lists of individuals affected to help them get discovered by hiring companies. There were approximately 2800 people across 40 companies in the dataset. During peak layoffs, this number would have been higher.
Please note: The only data extracted for this analysis was job title. No company or personal data was extracted.
A quick statistical calculation tells me the total sample of people I needed to analyse to represent the overall layoff population was 384, I have 7 times that amount with 2800. — skip this next section if you trust my statistics.
Sample size estimation
Where:
- Z is the Z-value, which corresponds to the desired confidence level - I am using 1.96 to represent 95% confidence level.
- p is the estimated proportion of the population (as this is unknown, I’m using 0.5 for maximum variability).
- E is the margin of error set at 5% (0.05).
- N is the population size of 500,000Note: I know the statistically savvy amongst you would have noticed the second variable I need to account for, ‘total number of companies analysed’. Yes I am short on the number of companies needed, but I am making an assumption that companies like Google, Meta, Microsoft, Tesla which had layoffs in the tens of thousands are likely to add a high amount of variability to the data and that if I was to account for those outliers, I’d probably squeeze the bell curve and the average number of layoffs per company would go down taking with it the variability and total sample I would need. Even then, I’m sure I’d still be short, but there’s nothing I can do.
Categorisation of roles.
Job titles come in a variety of flavours so with the help of ChatGPT, I categorised all the roles into Job Categories and Job Seniority. For the sake of brevity, I won’t go deeper into seniority beyond the graph below (graph 2) and I invite the reader to make their own interpretation of the data.
Analysis of job categories
So going back to the question that triggered this analysis in the first place:
“Which function in tech has been hit hardest in the layoffs?”
At first glance (Graph 1) it would appear that Engineering was hit hardest, at least in terms of absolute numbers. But given that Engineering is the largest department in most organisations, is this really that surprising? No is the short answer.
The question that we need to be answering is:
“Which department over-indexed against expected layoffs?”.
To answer this we need to normalise the data.
If we use Engineering as an anchoring point, we simply need to figure out the [Job Category] to Engineering ratio.
A couple of years back, I read an interesting article where someone had analysed the ratio of Data professionals against Engineers of over 50 tech companies. Luckily for me, that person has since updated the analysis and included Product as well. Using a similar search on Google I was able pull together the following ratios:
Product to Engineer ratio - 1:8
Data to Engineer ratio - 1:6
Designer to Engineer ratio - 1:9
Researcher to Designer Ratio - 1:5 (I converted this to Researcher to Engineer in my calculations).
Note: I couldn’t find a reliable enough number for the other departments for this analysis, therefore I have stuck to the core tech functions above. I would have liked to have included Marketing. If anyone has a reliable ratio for marketing against any of the above, please leave a comment with source link and I can update my analysis.
Results of Analysis
Now that we have our ratios estimated, we simply need to multiply the “Total Number of Engineers Laid Off” in my data set (898) by each of the above ratios to get the expected number of layoffs for each category.
A simple “Actual Laid of / Expected Lay offs” calculation also gives us the likelihood to be laid off for each function.
Based on my estimated ratios and my dataset we can clearly see that:
As a product manager, you are 2.6x more likely to be impacted compared to your engineering colleagues.
As a designer you are 2.3x more likely to be impacted.
As a researcher, you are 3.6x more likely to be impacted.
As a data person, you’re equally likely to be impacted.
How did we get here?
So how did we end up in a situation like this? Why are companies more likely to lay off employees in product, research, or design? I’ve given this a lot of thought and I keep coming back to trigger-happy leadership teams who failed to recognise the pandemic for what it really was—a bubble.
During the pandemic, many companies experienced a surge in demand for digital services and products, leading to rapid hiring sprees. With expectations of continued growth, they expanded teams across every aspect of the product, anticipating long-term demand. Features that would have once undergone thorough discovery, prioritisation, and sizing were now being churned out at breakneck speed. Product managers, once focused on delivering cohesive user experiences, found themselves reduced to managing feature factories, losing sight of the bigger picture. I believe this came at a heavy cost with product now facing a huge skills gap brought about by the lack of need to flex/develop core product development skills - but this is a post for another time.
When the world began to return to some semblance of normality, the anticipated growth slowed, leaving companies overstaffed, particularly in roles not directly tied to immediate revenue generation, like product management. This had a knock-on effect on design and research teams, which had been hired to support a legacy product structure designed for managing limited resources instead of the one the companies were facing.
The reckless expansion of leadership
Typically, pre-pandemic, employee growth was a function of sustained profitability or extreme necessity—i.e. "don’t hire until it really hurts." However, with companies literally having cash thrown at them due to an artificially inflated digital economy, caution was thrown to the wind. Leadership teams, eager to capitalise on the moment, built empires on unstable foundations, setting themselves up for an inevitable collapse. If ever there was a time when the saying "correlation does not equal causation" applied, it was during those peak covid years.
In contrast, engineering departments, which are essential to maintaining the infrastructure and product delivery, were somewhat insulated from layoffs. This doesn’t mean engineers weren’t impacted, but comparatively, roles tied to customer experience, product innovation, and design, which are often seen as more forward-looking, were hit harder when companies sought to reduce costs quickly.
You might be wondering, why wasn’t data hit as hard? I think part of the reason lies in the versatility of data skills. But more importantly, it’s because I believe product leaders still don’t see data as a core function alongside the traditional holy trinity of modern product management which includes: engineering, research, and design. Therefore hiring in data roles never over-indexed and as such couldn’t have over-indexed in layoffs.
A Reassurance for those affected
It’s important to remember that if you were affected by these layoffs, it wasn’t due to your skills or performance but rather a recalibration of expectations and staffing needs. Companies were forced to make decisions based on shifting market conditions, immediate financial concerns and pressure from VCs and boards.
In conclusion, while some functions in tech appear to have been disproportionately affected by layoffs, the reality is that these decisions were driven by external market factors and short-term thinking. As we move forward, I remain optimistic that the tech industry will stabilise, and those affected will find themselves in new and exciting opportunities soon. I hope we as a tech community have learned from our mistakes and the focus now shifts towards sustainable growth. I know I have, and I’ll certainly think twice before joining any company focused solely on growth at all costs—because nothing good comes from that.
Cautiously optimistic but positively skeptical.
Bhav
Bhav is the Director of Analytics and Experimentation at LeanConvert.