**Overall, deal activity is shrinking, not growing. Deal counts peaked in 2015–early 2016 and have fallen every year since. Typical deal size has gone up over the same period, and I see no reliable seasonality.** **Deal volume (the most reliable signal)** - Deals per year were about 929–935 in 2015, 993 in 2016, 687 in 2017, 309–310 in 2018, 111 in 2019 and 7 in 2020 (steps 3 and 4). - The two yearly queries differ slightly in how they handle the date text, but both show the same decline. - The monthly series shows the same shape. Counts rise from 51 in Jan 2015 to a peak of 104 in Jan 2016, then fall to the 70s in late 2016 and the 40s in late 2017. - By late 2018 only 7–18 deals a month were logged (7 in Oct and Dec 2018), and early 2019 shows just 3 per month. - 2019 and 2020 are probably partly an artefact of coverage. The last dates are 2019-12-20 and 2020-01-13, and 2020 holds only 7 deals. The data likely thins out toward the end rather than reflecting a real collapse to near zero, so treat the tail with caution. **Deal size (rising)** - The median disclosed amount per deal was about $1.5M in 2015, $1.0M in 2016, $2.25M in 2017, $3.9–4.0M in 2018, $12M in 2019 and $9M in 2020. - Fewer but larger deals suggests the market became more concentrated over time. Small deals may also be under-reported later in the series. - Only about 55–70% of deals in each year have a disclosed amount, so the medians describe only those deals. **Total dollars (noisy, no clear trend)** - Monthly totals swing widely and are driven by a few very large deals. Examples are about $2.75B in Aug 2017, $2.26B in Mar 2017 and $1.53B in Jul 2015, against typical months of $200–500M. - My yearly total queries disagreed. 2015 came out at $16.9B in one and $8.6B in another, because the amount strings were parsed differently. 2016 returned `nan` in one, so I would not trust the yearly dollar totals. **Seasonality** - No month stands out consistently across years. The 2015 rise from about 51–62 deals a month in H1 to 86–98 in H2 is part of the early growth ramp, not a repeating pattern. - Dollar spikes land in different months each year and look like individual mega-deals, not a calendar cycle. **Data quality** - A few rows have malformed dates: one is "15" for the year, and about 7 are blank. They are a negligible share of the 3,044 rows.