The market for exotic data (also known as “alternative data") is vibrant with datasets, ranging from detailed weather reports to Twitter sentiment to mobile app download demographics. Hedge funds and marketing teams use this data as a proxy for other data they want, but which might be harder to get. When you merge this data into Salesforce, it's even more powerful: these niche datasets can level-up your Salesforce instance to help your team with faster call prep, better lead scoring, and lower cost than brute force, "one size fits all" enrichment.
Below, I'll give an overview of the market for exotic data, and then focus specifically on the mobile data niche. I will conclude by sharing how we've seen major players in the mobile software market use exotic data to build incredible lead generation workflows.
Alternative data market overview
The alternative data market goes by two common names: "alternative data" and "exotic data". They mean the same thing, but you'll see them both out in the wild.
Speaking of wild, let's explore some of the common uses for "exotic data". Typical exotic data are niche datasets that cover a narrow topic, typically in great depth. For example, if you sold to high-net worth individuals, a database of "owners of tigers" might be an interesting proxy. However, we learned by watching the Tiger King series on Netflix that there is high variance in wealth in this data, and if you plotted the net worth of some of these individuals over time, you'd see high volatility.
Exotic data of this sort is not unusual. We're looking for correlation, not necessarily causation, and when you approach various analyses this way, it can lead to unexpected, even perhaps unusable results. But that's part of the fun. It's spinning a roulette wheel each time you dig and study.
To take the other term, "alternative data", for a spin, let's look at politics. In this example, your team may find that "Twitter users in the Bay Area who liked tweets mocking Kellyanne Conway" are good sales prospects. As you may recall, Conway served as press secretary for President Trump and famously coined the term "alternative facts" when describing attendance and the president's inauguration.
Social media is rife with huge datasets, and with the right access, you could mine it for reactions to this one story. You might discover that people who reacted negatively or positively to this story match a Twitter lookalike audience that has a historically high click-through or conversion rate for your business. Then, instead of using the lookalike audience, you can instead target Twitter accounts who liked certain tweets and drive ads their direction. You may find that your metrics improve. My point with this example is that there is a LOT of niche political data, and millions of ways to splice and analyze.
Types of alternative data
Alternative data covers every niche, and goes a mile deep into each one. That's the point. Aficionados of this approach love to explore, essentially spelunking through deep data-themed caves to seek out trends and insights that are hidden far beneath the surface. The payoff for these exercises is that you may be the first to discover a relationship, and you can exploit that asymmetry for profit. So let's look at the various types of alternative that exist on the market!
If you lived in California this summer, no doubt you were checking air quality data every day for some stretches of the summer. Historical numbers are tagged and stored for your data mining pleasure.
You can look at both the macro and the micro markets here. Real estate data is numerical, geographic, and even has ties reaching out to the bond and treasury markets. With homes being the most valuable asset on most families' balance sheets, it's a great indicator to master.
It may seem strange (it does to me), but there's a market out there for (anonymized) emailed receipts, showing all sorts of buyer history and intent. It's some of the best consumer behavior data available.
This is usually census-level data that gets combined with other datasets to make it more ingestible.
The secret's been out for a while, but people don't seem to mind — you have a mole on the inside. The inside of your phone, that is. It's pinging servers while you go about your daily business. So long as you keep your phone on and have any of the most popular apps installed, your location is getting harvested, packaged, and sold to the highest bidder. But hey, at least you can buy it too.
Less intrusive, perhaps, are the PoS systems that also track consumer behavior. They, too, will anonymize data (stripping out all personally identifiable info, like credit cards or email addresses) and sell it on the markets described in the next section.
Weather data is a great place to start. Weather events move markets; that much is proven. Being able to predict the weather, though, is where you can get a leg up.
Broadly speaking, any time some public deal is posted, it gets scraped and cataloged by a bot for easy use in data models. Examples might be insider stock trades, large government contracts, or real estate deals.
No introduction needed, but you might be surprised to learn that companies specialize in parsing news so you can use it in data models. News datasets typically include date published, publication, word count, sentiment, and known entities named. Most of these parsers today are driven by AI algorithms that do a surprisingly good job of "understanding" complex news articles.
If you watched Moneyball, you know how much love data nerds have for sport stats. Some data providers have aggregated all the stats, from all the leagues, over all the years. Have all sorts of fun figuring out if you can predict the S&P using the number of strike outs in the National League playoffs.
The movement of goods is literally what puts the economy in motion. Understanding tanker capacity, port openings, and how hurricanes impact trade routes can give you all sorts of insights into the whims of the broader and niche markets.
Our favorite alternative dataset; I'll go into much more detail about this little gem at the end of this post.
Google Trends have been studied ad nauseam, but don't let that hold you back. Tons of companies now are processing and repackaging this data to help you make the most of it.
Second to homes, automobiles are the next most common large purchase, and usually the next most commonly financed item (second perhaps to student loans). If you want to understand consumer behavior, it's helpful to look at what's happening to peoples' cars.
Finally, some providers have carved out a place in the market for loooong tail historical data. Interest rates going back to the 1100s? Yes, that exists. If you want to know how plagues have impacted the price of rice, you can dig that up too, going back to the Roman Empire.
Where to get alternative data
Are you salivating yet? I am. Or maybe it's just the bottom of my coffee mug staring back at me. It's such a pretty color. I never noticed it before.
Now that you're interested, intrigued, or maybe somewhat perplexed about the extent that you may personally be contributing to the mounds of data being gathered every second (whoops, there goes another terabyte!), I feel obligated to tell you where to get it. Here are a few solid sources to try:
Battlefin runs the largest online and in-person marketplace for exotic data in the world. Based in New York City, they match buyers and sellers of specific datasets, and allow the curious to browse the aisles (literally, in an expo center) of data providers and their mile-deep collections of spreadsheets you never knew existed.
It's not cheap, but neither are the best things in life. Or was it the other way around? Regardless, insights are priceless, so if you're into it, it's worth checking out.
What Battlefin is for hedge funds, Kaggle is for data science PhD students who don't have $5K burning a hole in their pockets. Want every email from the Enron affair? Look no further. And best of all, it's free. All of it. If you were in the market for data, I'd start here and then look elsewhere.
AWS Data Exchange
Is there a tech business that Amazon's not in? Seriously. Since S3 stores half the data in the world anyway, it only makes sense that Amazon would try to monetize a chunk of it.
Case study: mobile data
Okay, as promised, in this final section, we'll describe exactly how this whole alternative data thing works in the context of mobile data. If you're not familiar with mobile data, take a peek at the Data Points section on our Products page. It shows you every bit of data we collect and sell. It's a lot!
But why bother with mobile data at all? Well, some of the largest companies on Earth sell mobile products, and with over half of all Internet traffic today originating from mobile phones, it would be foolish to ignore it. You already know many top mobile brands: Google, Amazon, Apple, and Facebook. These companies alone represent several trillion in market cap, and they set global financial trends worth tracking. One interesting lens you can use to view them is the mobile dataset that MightySignal sells.
You may be wondering: how do you actually get alternative data into Salesforce? You can use an ETL (Extract-Transform-Load) platform like Xplenty, and we do recommend them for anything but mobile data (stay with me...) With Xplenty, you can hook up your data source, whether it's an S3 bucket, a SQL database, or a JSON file, and pull it into Salesforce using Xplenty's fancy data plumbing.
However, if you purchase mobile data from MightySignal, you can use our highly-efficient, direct Salesforce integration, which creates custom fields and objects, then synchronizes them automatically for you every week:
It literally doesn't get any easier than that! All you have to do is give us Oauth access so we're permitted to create the fields and write to them. From there, the sky is the limit.
Here are three quick patterns from real MightySignal customers that I can share.
Pattern 1: Incorporating a BI tool
In this case, they used Looker to combine and compute a lead score based on MightySignal and other B2B data providers. This custom datum got pushed into Salesforce alongside our direct Salesforce integration. Since our integration can't be customized, they needed Looker to compute the score, and a custom script to push that datum onto a custom Salesforce field that we don't touch. This gave the client the best of both worlds: a lead score that they can control and fine tune, and a set of robust mobile data that they can further use in Looker to refine and enhance their scoring algorithm.
Pattern 2: Incorporating marketing automation
In this flow, our client used Pardot to manage inbounds and marketing activities, like page loads and newsletter opens and clicks. They used MightySignal's API to pull some additional information about the lead, including which SDKs they were using, and based on that data, they'd use Pardot to push specific data into Salesforce. This is a highly configurable pattern that takes some time to set up, but then works exactly the way they want it to work.
Pattern 3: Incorporating an ETL
In this example, data stored in AWS Redshift gets mixed in with data coming from MightySignal inside Xplenty's ETL platform, and then pushed out to Salesforce. This makes the type of data pushed into Salesforce much more configurable, so if you don't like the fields we've hard-coded into our integration, then you can run it yourself through our friends at Xplenty.
To summarize, exotic or alternative data is a lever that top enterprise sales and marketing teams are using to gain unique insights into market opportunities. Whether it’s lead scoring or pre-demo research, using alternative data for any industry is a powerful way to get a leg up.
Mobile data is just one niche; there are countless others. Use our native integration or an ETL tool like Xplenty to ingest and transform alternative data from any source and push it into Salesforce, so the teammates who need it most can easily access it.
Oh, and if your eyelids are still open, thanks for reading!
This talk was originally part of Xplenty's Xforce Sessions on September 24, 2020. You can watch the original presentation here: