Sunday scaries: data influencers are out to get you | by Hugo Lu | Sep, 2023

“Data influencers” are the latest breed to unfortunately come into our industry. My old colleague once said “very few of the decent data engineers spend any time on Linkedin” — he was right.

Hello ? I’m Hugo Lu, a Data Engineer who’s also worked in Finance and now CEO @ Orchestra. Orchestra is platform for managing data release pipelines that helps Data Teams release data into production reliably and efficiently. I write about what good looks like in Data.

The influencer marketing, err, market I guess is forecast to be worth c.$20bn this year. That’s a lot. Historically, it’s made a reasonable amount of sense for me since the money-making impact of Influencers (and social media in general) plays on extremely powerful emotions; a fear of missing out, a feeling of jealousy, a desire to be like someone. Good for making people buy stuff and therefore making money. These are all documented in Rob Cialdini’s book, a leading expert on psychology and persuasion in particular.

Indeed, most of this market found its nascency in Instagram, but really the original “Influencer” market was probably celebrities being paid to promote stuff on television or radio. Instagram, and other social media in general (and I stress social) is merely an extension of this.

In the workplace, where we work, where data engineers are relevant (there are no famous data engineering accounts on Instagram) the relevant forum is unfortunately Linkedin, technically, a form of social media — it’s more a networking tool / marketplace in my mind, but we let this slide for now. I exclude youtube, because while there are many “influencers” on youtube too, I think there are actually some really solid videos out there. The first time I started using Docker for example, I found TechWorld with Nana and like it’s legit. It’s literally hours long videos on how to do something relatively difficult. She has 900k subscribers. Filming a 3 hour tutorial on Docker is not something I would wish upon anyone, so fair play.

Back to Linkedin Influencers — within this bucket, there are many types. There are “Career influencers”, “Finance influencers” and “Data influencers” (non-exhaustive). Data influencers have many thousand followers on Linkedin, which means there are many thousands of people that the Linkedin algorithm shows their content to every day. They post about a variety of topics, from how to be good at doing data stuff, to trends, to “thought leadership”, to shilling things like educational courses and other people’s SaaS products.

I’m not huge on influencers generally, and I am really not huge on them creeping into my line of work. Let’s dive in to why.

In economics, if you’re willing to buy something for a price, there is an explicit assumption you get something for it. If I am willing to buy a can of coke for 50p, it’s assumed I’m going to be at least as well off as I was before I bought the can of coke after I buy the can of coke. Maybe it’s hot like today (32 degrees) and I really don’t care about the 50p, I just need a cold drink and it makes me feel good. Pretty intuitive.

Influencer marketing preys on emotions. It changes your desires. Perhaps I didn’t actually think I needed a can of coke, but I go on my phone, see a video of David Beckham drinking some Pepsi and I’m like “shit I really wanna be like him, he looks so cool, I really want a can of pepsi now”. So I go and get some pepsimax or whatever and I’m then loads better off.

Apart from — my desire for pepsi literally came out of an ad. It literally came out of nowhere. If I had never seen the ad, I would’ve been fine without a pepsi. Now, I kinda have a desire I didn’t have before satisfied but I’m also 50p poorer.

Another good example is nicotine / smoking. You have a desire, you pay money to satisfy it. But the desire you have is one you’d rather not have. You’re not better off after you buy the cigarettes (in fact in this example you’re worse off because cigarettes are also bad for you as they cause cancer).

This is kinda how I view Influencer marketing. There’s a load of stuff I don’t care about and I am happy to not care about it. You see an ad, it makes you feel something, you get a desire, and you pay money to satisfy it. At the end you’re not really better off, but you’re probably non-negligibly poorer.

This is a necessary evil. We are all human, we all feel things, and we can all be manipulated. If we’re just rolling around being manipulatable then you can rest assured there is a market for that. It exists necessarily. That market is called advertising. Instagram is the perfect setting.

In data, it’s a bit murkier, and it’s also not on Instagram mainly — content is on Linkedin.

Sometimes, people just post for the sake of it. They like the attention. They don’t ask you to buy anything but they enjoy being listened to.

Other times, they’ll use Linkedin as a funnel for other things e.g. Youtube. They make money from Youtube. Here, you pay, not with money, but with your time.

Then there’s the VC angle that’s out as a bit of a conspiracy theory on the reddit threads linked above. The idea is, if you’re a data (or any workplace) influencer, you can convince a small proportion of your 100k followers to trial something. This has value.

This has value because if I’m a start-up, and suddenly I get 100 people sign-up over the weekend, it looks good. It doesn’t matter that the likelihood any of them become real customers is low, because no-one knows that. No-one knows the traffic generated from Linkedin is actually super low conversion traffic. Just show a graph of cumulative signups to a VC and you’re golden.

There are, therefore, two tracks for a data influencer to make money.

The first is to become one, and then pivot to running a product business / a business where they can sell stuff. This requires them to actually curate their audience, because if people don’t convert from free trials or actually part with cash, the influencer doesn’t make any money.

The second:

Become an “Advisor” to a start-up. This means directing as much of their audience’s attention to this start-up as possible. The start-up compensates the influencer with shares, and the start-up gets sign-ups.

Sign-ups are valuable for the reason above. Not because they are likely to lead to revenue but because having sign-ups is seen as an intrinsically desirable trait to many VCs. This means start-up owners are willing to pay for it.

It also spells disaster for the layman like you or I who wanders onto Linkedin one day and is bombarded with “Data is Dead”, “the hottest thing you don’t know about” or even the dreaded “Statement plus three dots on a new line” — basically boring, baseless, same-y content.

If there’s no incentive for people with many followers on Linkedin to post about anything useful then Influencers will:

• Talk about trends authoritatively to grab your attention, not to speak about relevant trends in your industry
• Recommend tools in a biased way
• Make you go to industry events you may not actually find beneficial
• Influence how you go about setting career goals which may be completely unrealistic and actually be damaging to your career in the long-term

You get the picture. Every time you read these posts, it’s like watching Beyonce, Pink and Britney drinking Pepsi Max in the Colosseum . You’re not actually getting any new information you care about. It’s just giving you disinformation / random fictional content that trigger emotions that make you do stuff the “sensible” you wouldn’t have done. Helpful information is like an actually good Linkedin post where you learn something factual you didn’t know before.

Most of the stuff in Linkedin is like the dude on the right. There are very few people who consistently post factual information (or even better, maybe a link to a medium blog where there’s actually enough space to learn something instead of a 300 word linkedin post) — which I would equate to the ad on the left.

Every time you read one of these posts, you pay. You pay in a few ways:

• The time it takes you to read the post
• The impact on your decision-making of having disinformation
• The impact on your decision-making of having a desire to realise the [insert new hot data trend here] for your data team you didn’t have (or need) before

“Data Influencers” are winning because:

• They get compensated for followers and for writing generic and generally unhelpful bits of content with complicated flowcharts generated using random bits of draw-it-yourself software

The community is losing because:

• The noise being made in the industry does not necessarily align to best practice / moving it forward
• There is a lot of noise, which makes decision-making for us (data people) hard (one can have too much information)

So what do do?

I once told a person in my team (very smart data engineer) about something I had learned on Linkedin and they simply responded with:

None of the good engineers spend time on Linkedin

And I was kinda inclined to agree. The most competent people (and this goes for anything really) are probably out there achieving and enjoying their lives, smashing through work in record time so they can go and enjoy the free time they just earned. They don’t spend that on Linkedin — there’s only a small proportion, the tip of the iceberg, that do.

The best thing you can do, therefore, is get (tf) off Linkedin. If you’re aspirational, and aspire to do as the best do, the best people are probably elsewhere. Maybe they’re at conferences meeting people, maybe they’re reading books, maybe they’re running side-projects with friends and learning by doing. But they’re definitely not loitering on a landing page for an online thousand dollar data engineering course (and if they are, they’re probably the ones selling the content) with utm_source=linkedin like you.

I appreciate there are lots of premises that underpin this argument that could feasibly be denied. Anecdotally, and from personal experience, it feels like most of them are probably true. However, it would be good to validate them a bit more formally.

To that end, please hit me up with suggestions. Right now I was thinking of doing some webscraping of Linkedin for posts from specific influencers. You could then cross reference the posts with things like average length, frequency of lists occurring, check the content vs. chat GPT and calculate a misinformation score, compare posts to other posts to see how generic they are, do some sentiment analysis on them etc. etc. to literally demonstrate that the content we’re discussing is, indeed as I’ve argued here, not that helpful. I’ll hopefully do this at some point but for now, over n out ?

Side-note

You might have realised that I’m the CEO of a data company and also regularly post on Linkedin. I try to keep my content as non-generic as possible. I want to provide helpful information on best practices in Data. Having actually been someone who “fell into data” (as most of us do), I was a Head of Data, I know a fairly good amount of data engineering (see the rest of this blog) and I’ve written a lot of SQL, which I like to think lends me some credibility! I’ve never worked at FAANG, but it seems that’s made my views and experiences more representative of people in the data diaspora. I try to curate an audience of data engineers, analytics engineers, and heads of data who are open to learning more about what best practices look like in data — not 18 year-old masters students who want to become staff engineers at FAANG. They may end up needing Orchestra, they may not. I’m very interested in both groups as I think we can genuinely add some value by sharing information on how to achieve best practices in Data. So please feel free to connect and keep me honest ?

References:

1. https://influencermarketinghub.com/influencer-marketing-benchmark-report/#toc-2