Hello, welcome back to Papers with Backtest podcast. Today, we dive into another algo trading research paper. We're looking at the growing world of alternative data and specifically, yeah, the insights we can glean from information scraped directly from the web. It's a fascinating area, definitely. And it really is. And, you know, recent research indicates just how significant it's becoming. I saw something like a full 50% of institutional investors. Yeah, that's right.
Half of them are planning to ramp up their use of alternative data pretty soon. Wow. It's a striking statistic, isn't it? It really underscores the shift we're seeing in how investment decisions are getting informed. Absolutely. And, you know, all the different types of alternative data out there, web scrape data, well, it seems to stand out as particularly popular. Exactly. So, okay, we know it's popular, but when we talk about web scrape data. What exactly are we referring to?
Like, what does that actually entail in practice? Right. Let's unpack that. Essentially, it's data that's automatically collected from publicly available websites. Right. Just out there on the internet. Exactly. Specialized companies use programs, bots basically, to periodically harvest this information. Okay. So they're just constantly scanning and pulling data? Pretty much. Or sometimes they can get it more directly through public APIs. Ah, application programming interfaces. Yeah, exactly.
Which let computer systems talk to each other, share info in a structured way. Makes it a bit cleaner sometimes. And there are companies that specialize in this. Oh, yeah. You've got vendors like Quandl, Saver, Thinknum, Yipit. Quite a few operating in this space now. Gotcha. What's really fascinating here, though, is just the sheer volume of information we're talking about. Right. The internet is just vast. It's incredible.
Something like, what, 4 billion web pages containing around 1.2 million terabytes of data. That's hard to even comprehend. It is. And hidden within all that is potentially, well, a treasure trove of signals for investors, if you know where to look and how to interpret it. Absolutely. So let's think about the kinds of data you could pull. Job listings, maybe. That's a great example. Like a company that's aggressively hiring. That often means they're growing, right? It just makes intuitive sense.
It does make sense, generally. More hiring often signals. Expansion, investment, good things. But I guess, could it also signal something else? Like maybe problems? High turnover or restructuring? That's a really crucial point. Context is absolutely key with any data, but especially alternative data. An uptick in job postings alone, well, it might not tell the whole story. But if you start to layer in other information, then the picture can become much clearer.
Okay, like what kind of other info? Well, take company ratings, you know, on sites like Glassdoor. Ah, right, employee reviews and ratings. Exactly. So if you see a company's ratings going up, getting better, especially if that's happening at the same time they're increasing their job postings. Oh, okay, so hiring and people are happier there or getting happier. Precisely. That could be a really compelling indicator, positive momentum, maybe a sign of a healthy, expanding organization.
That makes a lot of sense. It's about connecting the dots, looking at multiple signals. You got it. And it's not just about like internal company stuff like hiring, is it? You mentioned online retail data earlier. Yeah, that can be incredibly telling too. Think about e-commerce sites. Okay. High product rankings, you know, best sellers. Yeah. That often suggests strong sales. Sure. If something is flying off the virtual shelves. Exactly.
But conversely, if you suddenly see a lot of heavy discounting on products. Ah, yeah. Big sales, price cuts. That could be a red flag. Might point to weaker demand, maybe inventory buildup they need to clear. Okay, this is starting to paint a picture. But maybe, maybe a concrete example would really drive it home. This is where I started to see the real potential. Yeah, definitely. There's a really good example focusing on Amazon back in late 2018. Amazon, okay. What happened? So picture this.
It's December 26th, right after Christmas. Amazon puts out a press release. Announcing? A record-breaking holiday season. They were quite specific, too. Oh, yeah. They said the Echo Dot was their number one top-selling product across all categories on Amazon. Wow. Number one overall. Not just devices. Exactly. And that the Echo Dot and the Fire TV Stick weren't just the top Amazon devices, but they were the best-selling products, period, across all of Amazon for the holidays. That's huge.
And they also mentioned millions of Prime members using Alexa for voice shopping, GIFs, devices, everyday stuff. Okay, so a big announcement. What happened next? Well, here's the kicker. Amazon stock price, it surged massively by about 23%. 23% just from that announcement? Seems like it. And crucially, during that same period, the S&P 500 was essentially flat. Whoa, so the broad market wasn't doing much, but Amazon shot up.
Right, which strongly suggests that the market hadn't fully priced in just how successful Amazon's holiday season had really been. It was news. Exactly. It sounds like news. Maybe not to everyone. This is where the alternative data comes in. This is exactly where it comes in. For those tracking it, this surge might have been, well, less of a surprise. How so? Think at them. One of those vendors we mentioned. Yeah. They were using web scrape data, specifically from Best Buy's website.
Okay. Tracking Amazon products sold through Best Buy. Precisely. And their data actually showed the increasing sales strength of Amazon products, like the Echo devices, throughout the holiday period. Starting way back on Black Friday. So weeks before the official Amazon announcement. Weeks before the data was already indicating strong performance. That's the information advantage right there. It really can be. And it wasn't just some vague signal about Amazon products.
No. Best Buy's website data also seemed to corroborate Amazon's specific claims. It showed the Echo Dot and the Fire TV Stick as top sellers in their respective categories on Best Buy's site too. So independent validation from a major retailer's own data, that adds a lot of weight. It really does. Makes the signal much more robust. Okay, that's compelling. But you mentioned job listings earlier, too. Did that play a role here? It potentially adds another layer, yes.
If you look back throughout 2018, before the holidays, the number of open job positions related to Amazon Alexa on Amazon's own corporate careers website, it saw a remarkable increase. How much? About 53%. It went from around 500 postings. to over 750 by year end. Wow, a more than 50% jump in Alexa related jobs? Right. So for investors tracking that data point throughout the year, it suggests a major strategic push, significant investment in the whole Alexa ecosystem.
Which might make you think, hmm, they're betting big on Alexa. Maybe the related hardware sales will be strong too. Yeah. Especially coming into the holidays. Exactly. It could indicate an expectation internally of future growth, product expansion. Making those strong holiday sales for Echo devices seem more plausible, maybe even predictable to some extent. That makes me wonder, though. OK, we see this this link potentially between more Alexa hiring and strong Echo dot sales.
But does that automatically mean it's, you know, great for Amazon's overall profit? And that that brings us to a really important point about the limitations. Right. Because Amazon is huge, diversified. Exactly. So while the sales data gives us great insight into how specific products are doing. and the hiring data suggests investment in a specific area, it doesn't instantly tell you the whole story about the bottom line for a giant like Amazon.
A hit product line is great, but maybe it's still a small fraction of overall revenue or profit. It could be. And similarly, those increased job listings, yes, they can signal growth and investment, but they also point directly to potentially higher costs, right? Yeah. More salaries, more overhead. Which could actually hurt profitability, at least in the short term. It could absolutely affect profitability. So you need to consider that side, too.
So it's definitely not a magic crystal ball, this web-scraped data. It doesn't just give you all the answers on a plate. No, absolutely not. It has its limitations, just like any data source. You have to be aware of them. So what's the takeaway, then? Is it still valuable? Oh, I think it's incredibly valuable. The real power, though, often comes when you combine it. Combine it with? With traditional financial data, the earnings reports, the balance sheets, And also with, you know.
Good old fashioned qualitative analysis, understanding the business, the management, the competitive landscape. So it's another piece of the puzzle, not the whole picture. Exactly. It's about building a more complete, more holistic picture of the company and its prospects. Okay. But even with those limitations, it seems clear that web scrape data can provide really valuable and maybe crucially timely insights. That's a key word. Timely.
It can offer a different lens, maybe a quicker lens, to view a company's performance and what might be coming down the pike. Giving you potentially an early peek behind the curtain. Yeah, something like that. Identifying trends, maybe potential turning points before they become widely apparent in the traditional financial reports that come out quarterly. So it encourages a more dynamic, more real-time understanding of what's happening. I think that's a great way to put it.
It's another tool, a powerful one, for staying informed in a fast-moving market. Thank you for tuning in to Papers with Backtests podcast. We hope today's episode gave you useful insights. Join us next time as we break down more research. And for more papers and backtests, find us at https.paperswithbacktests.com. Happy trading!