Famously described as ‘soft, sad freaks on an unprofitable website‘, a description I think most of us took as a compliment, Tumblr users are highly resistant to being advertised to. It’s one of the reasons the site has avoided becoming exactly like every other social media platform — and one of the reasons it’s stayed unprofitable, although the current staff strategy of convincing users to pay for utterly meaningless upgrades like covering their dashboard in small crabs seems to be a winner.

As such, it’s not the kind of platform that’s well-suited to branded accounts trying to set themselves up to advertise to their followers. Tumblr users spot that a mile away, and have no interest in buying what you’re selling. Tumblr’s a community focused website suited to chaos, largely populated — at least in my circles — by broke 20- and 30-somethings. Mostly postgrads, in my experience, although that’s probably a side effect of hanging out in the academic corners of the website.

It’s therefore not a website I’d recommend joining if your aim is to advertise yourself as an author. But as someone who has been on Tumblr since 2012, met some of their best friends there, and feels less alienated by it than by every other social media platform, I wanted to know if I could use it effectively for authorial advertising.

I knew, first of all, that my posts on Tumblr had convinced some long-term followers to buy my book, and also that a few people had followed me after reading the book, or had looked it up because of one of my posts. That suggested there was an audience on Tumblr who would find The Butterfly Assassin interesting. But while I’d been doing my best to reach that audience the organic way — reblogging relevant posts, sharing information about the book with my own followers, etc — it’s hard to get the traction needed if you don’t have a particularly substantial follower count.

What I needed, I decided, was a way of getting my posts in front of the eyeballs of more Tumblr users. Enter: Tumblr Blaze.

For those unaware, Blaze is essentially a promoted post — pay $10 and they’ll show your post to 2,500 people. You can pick a country to target, but otherwise it’s completely random, because Tumblr doesn’t collect sufficient personal information to target ads to more specific demographics (which is one of the things I like about it). When it was introduced earlier this year, users embraced the opportunity with glee, using it to inflict on random strangers everything from shitposts to pictures of their cats to links to their fanfic.

I decided to Blaze a post I made a few months ago that hadn’t had much engagement. It features an image with some nice quotes from Goodreads review, and then a little information about the book in the caption, with buy links. Crucially, it’s focused on alerting people to the fact that the book is only 99p on Kindle. This meant I wanted to focus my Blaze post to the UK, since that’s where the price drop applies to, and I hoped that the deal might mean it caught people’s eye.

Blazed posts at this tier run for 24 hours, during which time they’re shown to a minimum of 2,500 people. If they reach that threshold before the time is up, they keep running. Once the 24 hours is over, Tumblr sends you a report with the statistics, so you can see how much of a difference Blaze made. (There are higher tiers, but I decided to stick with the cheapest for this experiment, because I am one of those broke 20-something Tumblr users myself.)

Well, the statistics for this post weren’t particularly inspiring. Although the post had garnered 3,600 impressions in that time, it only received 19 likes and 3 reblogs. True, that’s about 20 people who now know about my book who didn’t know about it before (I didn’t recognise the usernames of all except one of the blogs who engaged with it), and it’s possible that others who saw it in passing would have had their interests piqued and might buy it if they saw it mentioned again. But as far as I could tell via the affiliate links in the post itself, and via imprecise stats like Amazon rankings, I didn’t actually sell any books.

Out of interest, I decided to Blaze a completely unrelated post, to see whether this was a normal level of engagement to receive from Blaze or whether it was because it was a clear advertisement, and Tumblr users hate having things sold to them. The post I chose to Blaze as my ‘control’ in this experiment was a text post I had just made about language learning:

I chose this post for a couple of reasons. Firstly, it was a reasonably sincere post, not a joke; Blazing something absurdist and strange might tap into the Tumblr psyche, but it wouldn’t be as useful as a point of comparison. I wanted something that had a similar tone to my other posts, but was very clearly not trying to sell something. That was the second reason I chose it: there’s no way you can slice it that makes this sound like I’m trying to sell anyone anything. It’s obviously just a piece of personal experience framed as advice. And thirdly, the same post had received reasonable engagement on Twitter when I posted it there, so it was obviously the kind of thing that resonated with people if they saw it.

Thus, what I was testing here was primarily how effective Blaze was at getting my posts to a wider audience, and whether advertisements did worse than regular posts.

I immediately noticed a difference. Not only did my post start getting likes and reblogs within minutes, but they were definitely from Blaze, not just from ordinary browsing (tagging it as langblr would probably have helped). I know this, because one user added tags responding positively to the fact that it was Blazed:

Honestly, it made me feel faintly guilty that I’d blazed the post as part of an experiment, and not purely out of fond feelings for fellow language-learners! But it showed that Tumblr users respond positively to non-ad usage of paid posts more than they respond to posts they see as benefiting only the person who blazed them. If they think you’re doing it because you want them to get something out of it, they’re much more likely to enage with it.

Engagement slowed as the hours passed since posting (and since the Blaze campaign began), but it still continued to trickle in. By the time my Blaze campaign ended a day later, the post had received 77 likes and 16 reblogs, and the report was as follows:

Not only was overall engagement significantly higher, but the number of ‘earned’ impressions and engagements — e.g. the number of people who saw it not because I’d paid for them to see it, but because somebody else who’d seen it via Blaze had chosen to share it the usual way — was also higher, meaning that the post escaped paid containment and began circulating organically. Since the Blaze campaign ended, it has continued to accumulate engagement, because it has now begun circulating in relevant language communities: reblogs beget reblogs.

This was a deeply unscientific experiment on various levels. My book post was old, and had already received as much engagement as it was going to get naturally. My language learning post was new, and still had the potential to be picked up by other study blogs, even without being Blazed. The first campaign went live at 11.30pm; the second just before 5pm. (This wasn’t within my control, as it depends entirely on how long it takes the moderators to approve a campaign after you pay for it.)

But the findings seem to correlate with what I expected: people on Tumblr are highly resistant to posts where the main purpose is to sell them something, even if it’s something that they might be interested in, and reaching a larger audience means nothing if that audience doesn’t want to engage with it.

In other words: if you’re looking for a website/platform where paid ads will be a good return on investment, Tumblr isn’t it. Even if you’ve been there over a decade, like me — after all, the users who see my Blazed posts won’t know that, and to them I’m just another person trying to convince them to part with their money.

I don’t have any conclusions to draw about the value of Blaze for any purposes — frankly, I’m glad Tumblr is giving its user base ways to give them money directly if it means they don’t have to pivot to selling all of our personal data. But it seems clear to me that if authors want to make use of Tumblr as a platform, then the best way to go about it is the old fashioned way: be a person. Talk to people. Share posts, your expertise, and your opinions; join in with discussions, and become part of the community. And then, maybe, people will buy your books because you’re their friend and they trust your taste and judgment and opinions about storytelling and they know they’ll enjoy it. But even if they don’t, you’ll have made friends, and created a space for yourself online that isn’t about the perpetual hustle to market yourself and be a brand.

I probably could have told you that without spending $20 to experiment with paid posts. But sometimes it’s worth doing science on a question to make sure what you thought was the answer actually holds up to scrutiny, and you know it’s science, because I wrote it down, and that’s the part that matters. I hope we both learned something from this experiment, even if it didn’t sell any books.

And if you’re looking for a space online to embrace unprofitability, anonymity, shenanigans and a whole lot of puns, and you’re not already on Tumblr, maybe now’s the time to join.

But in the meantime, maybe you would like to buy my book… after all, it’s only 99p, or the paperback makes a great Christmas present…