Big data. If you've spent some time online or anywhere in modern society within the past couple of years, you've probably heard the term thrown around. If you're involved in communications in the crypto space, you've probably also heard the dangers of "users as the products" and how we are commoditized for our data.
As technology continuously pushes the envelope of what is possible, the data this technology can be fed is increasingly valuable. One of the premier metrics through which this data can be employed is through the concept of predictive analytics. Predictive analytics is a process in which computational resources analyze and interpret data as a means to attempt to answer questions it is given. For example, a large company may want to know how their stock and sales may be affected if they broadcast a highly controversial advertisement. Nike and Gilette didn't just take a gamble, they crunched the numbers and determined they'd be better off by taking a political stance in their advertising.
It's not a secret that if companies or individuals can create a system that predicts the future for them, the potential payoff can be monumental. And this goes beyond profits- banks and hospitals can employ predictive analytics to determine weak points in their infrastructure, protecting their clients and patients from data theft. Colleges and universities can have a better understanding of what their incoming class of students might look like. Any question that fundamentally could be solved given proper data, theoretically can be solved through this budding industry. You might be asking yourself, "Could this concept be used to forecast future prices of coins?" The answer is yes, and it's readily available with Endor- but we'll talk about that in the next blog.
The problem with predictive analytics, in the reality of today, is it sucks. If you aren't Google or Facebook, and thus don't have access to massive amounts of proprietary data, it's barely even worth your time. The college I attend, for example, employs predictive analytics to, like I alluded to previously, build data on the upcoming admissions cycle- how many kids will put down deposits and what their profiles will look like. This initiative is 26% accurate. In other words- useless. By and large, this is the case for the early adopters of this concept. Despite spending tens of thousands of dollars a year to get access to resources offered by IBM or AWS, the material output is arguably 0.
There are three main obstacles to further predictive analytics adoption: it's expensive, it's inaccurate, and it's slow. Endor Protocol is such an important project because it solves all three flawlessly. This isn't theoretical, Endor is live, being used by dozens of small and medium businesses (SMBs), and by enterprise clients such as Coca Cola, Walmart, and Mastercard.
The two key components to Endor are its utilization of blockchain and social physics. The main reason why predictive analytics is so expensive for users is because those who provide the resources split said resources among each user. Two restaurants could be using IBM platforms to ask the exact same questions, but IBM has to use their resources to answer that same question twice. Multiply it worldwide and it makes sense why the average cost of a predictive question exceeds $10,000.
By building the protocol on Ethereum, resources are never split, always shared. Users pay-as-they-go, sending EDR attached to their question proportionate to the amount of resources required to answer it. Through this metric, Endor gets cheaper the more it is used. In a beta late 2018, it was found that the average cost of a predictive question was just under $100. This is over 99% cheaper than the industry standard, which exceeds $10,000. The long term goal of the team is to drive that cost down to $1 per prediction as more and more people use it.
In the same vein, the blockchain nature of the protocol rapidly reduces the time it takes to return an answer. In the beta, answers were returned in hours, versus months on average across industry competitors. However, blockchain only partially explains the increase in speed. The other half of the equation is Social Physics.
Part of the reason why predictive analytics is so slow and inaccurate is because it operates on zero assumptions. This is silly, argues proponents of Social Physics, because humans are predictable, and assumptions and predictions can be made off of their tendencies. A great example of the benefit of this can be seen through one of Endor's beta uses: detecting the how exchanges volume and activity may fluctuate. By looking at the accounts and their tendencies interacting with the exchange, it can be seen quite accurately which users will move to new exchanges, become "whales", and so on. Perhaps the most important metric of that beta was that accuracy was improved ten times beyond industry standards.
The result is 99% cheaper, 99% faster, 10x more accurate predictions. Again, this is all already in use. Next post I will show how I am using Endor Protocol every day to receive price predictions on coins for guaranteed profitable trading.
Results of the beta: