CRYPTO: Did ICOs raise $300 million or $2 billion in September? Depends who you ask.

Last week we published some grim ICO figures, which made the rounds in the media, suggesting that token offerings are 90% down on a monthly basis relative to the peak. We were challenged on our figures based on two sources: Elementus and Coinschedule (ICO Rating is another great reference). While our number floated down to $300 million, some of the others saw September as $1B+. As an aside, we want to show the largest number possible to frame the best story for a delicate and growing space. This is why we began adding venture capital equity investment into crypto companies. When looking at that particular chart, our trend is at over $1 billion in August. So let's explore the delta.

First, there are some chunky and problematic figures which we chose to treat differently. For example, CoinSchedule lists Rubi-X as a $1.2B ICO entry for the month of September, which we have not been able to verify otherwise and chose to exclude. The Venezuelan Petro ($700MM+) we also ignore, as it is at best a government-backed monetary unit, and at worst an experiment in sovereign fraud. Second, there are various timing differences. Take the $134MM into tZero, which we had already accounted for at announcement in July. Lastly, taking a closer look, many of the ICOs we chose not to include are self-reporting a "completed" ICO, and then a data spider is taking their softcap as the amount raised. We generally exclude data where the confirmation of a meaningful raise ($10MM+) is hard to pin down.

Further, Elementus tracks monthly flows as they happen. This means if an ongoing ICO is 40% through its time period, they will have counted accrued fund flow. We track data at period end, meaning that only closed offerings are counted. Such an approach will not give credit for capital in flight, and perhaps there are good months ahead if indeed flows are strong. But this methodology difference should generally wash out, unless large chunky and unusual things are happening (e.g., Telegram and EOS). 

Our final point on this is to revisit our data sources. We focus on analysis, and leverage other primary sources for much of the underlying gathering, which we then scrub. In looking at the process, we counted over 30 ICO trackers used in our aggregation process. The kicker though is the short half life of the sources -- charted below. We find ourselves swinging between various sites and their increasing and decreasing data quality! So if you ever think there's something we should pay attention to, do let us know. And without further ado, here are the adjusted figures, telling the same story as before.

84d1a718-d78b-4bd8-a684-831b0c1b77fa[1].png
918280df-100d-4f27-a0fb-1c91d562fda4[1].png
19feeb78-6581-40bb-a43e-956c09af9885[1].png
2dcc0052-c813-4cc9-81d5-9ccca37bd7d2[1].png
47529b13-b0c7-41f3-9d18-80e49ddf8af5[1].png

Source: Cryptoglobe (ICO data reporting), ElementusCoinSchedule