Marketing Stats - Web Advertisement on Twitter

5 years ago | Patrick Rochefort (Member)

Rounding off my analysis on prior posts about Facebook advertising and Reddit advertising, here's my next analysis, this time on Twitter.


Budget: $35.00 -- Timespan: 7 days.

The first thing out of the gate that I REALLY liked about advertising with Twitter was the very comprehensible targeting. Of course you had your usual demographics analysis, but there was some spectacular other options, like: "Advertise to your followers" and "Advertise to the followers of the people you list here", the latter option was FANTASTIC for targeting. If you know the twitter handles of, say, five or ten authors who write things like you do, targeting their followers means that you're probably planting on fertile ground.

The next thing that I liked was that Twitter encouraged you to have more than one tweet as an advertisement. If you want a great experiment in honing your pitches to 140 characters and below, this is a fantastic exercise. I created four tweets that advertised the story, and saw some very different response rates from each of them, which I'll detail further down this article.

Let's start with the raw stats, in a big image file right here.

Overall, $35.00 bought me 43.8k impressions, with (27/92)* clicks. Why the asterisk? 27 of those clicks were directly to the site, while the total of 92 included people who went to my Twitter, clicked on the website card, favorited, or retweeted. Twitter does some pretty good granularity that way, but they don't always explain it well.

Overall click-through rate: 0.06%. Devastatingly low. This is one-fifth the rate that Reddit offered me. On the upside, I served a lot of impressions, but overall it's clear that on Twitter, people generally don't want to click ads, even less so than they want to on other sites.

Engagement rate: 0.22%. Not terrible, here, when we factor in people engaging with the tweet or otherwise interacting with it. As we've covered in prior entries, 0.2-0.4% is the average you'll find most non-targeted advertisement fall into. The fact though that this was targeted advertisement, is another black mark on Twitter for adverts.

And finally, our almighty Cost-per-click: $1.30. Kinda pricey. I'm paying Facebook cost-per-clicks for a response rate that's considerably lower. The impression rate is pretty awesome, which is nice for building brand awareness, but overall if people ain't clickin', they ain't buyin'.

I wouldn't use Twitter again unless my goal was new brand establishment, or to bring in a wave of new eyes and awareness on a story product.


By the tweet:

So this was a part of Twitter advertising that was really, really valuable for me. Finding out which tweet I wrote had the most engagement, and drew the most interested eyes to the story.

My top performer, at 0.25% engagement rate: "Does the necromancer who butchered her husband and son deserve a Detective's justice, or a mother's revenge?"

In retrospect, it's an obvious choice. It's powerful, engaging, leaves the reader with a compelling question, and people clicked through a little more there than any other one. As a result, I've included the line in other marketing and synopsis of the story since.

Tied for 2nd place, at 0.22%:

"Heather Blackthorne once hunted down necromancers, until one hunted down her family. Now he's come hunting for far more."
"From Winter's Ashes - A Detective with nothing left to lose, against a Necromancer with the world to gain."

And my last-place finisher, in 4th, at 0.19%:

"In a world where everyone possesses magic, a Detective seeks justice on the Necromancer that butchered her family."


In conclusion:

Twitter, at this expense level per engagement, is probably prohibitively expensive for a webserial's conversion rates. The ability to finely target other authors followers is very nice, but is countered by the fact that Twitter users are particularly ill-inclined to engage with advertisement there.

The only uses Twitter will have for a webserial's economics may be starting up a new brand/story, getting people interested in a new book or e-book release, and generally getting your brand name out there. If so, I would recommend including your name in the tweet, so that even if the title doesn't stick, your name hopefully does.

I won't use Twitter again for From Winter's Ashes, I think, but if nothing else, the $35.00 spent was valuable to see, in start stastical payout, what one-line summaries of the story were more effective at hooking people.

From Winter's Ashes: A Detective with nothing left to lose, against a Necromancer with a world to gain.

Read responses...


  1. Wildbow (Member)

    Posted 5 years ago

    Hey, Rochefort - nothing to add except I really like the way you presented your findings. The line at the end about the specific scenario where it'd be useful in particular.

    I can say that I started Worm and started at 0. I built up to 25,000 readers or something by the end, then moved on to Pact, and only ~4000 stayed with me. Built up to about 5-6000, ended Pact, moved on to Twig, and only about ~2500 moved over with me. Now I'm building up again.

    I'm retaining that core audience of readers; the really loyal ones who pay my wages. Even with the smaller number of readers, my average income is steadily rising - I'm not in dire straits, but those lower reader numbers can be frustrating. Getting/finding that audience and telling your readers where you're relocating to is something of a challenge, and perhaps twitter is the medium for answering it, in part.

  2. ubersoft (Member)

    Posted 5 years ago

    I actually saw the detective with nothing to lose ad in my stream and was mildly curious. That's more impressive than it sounds because I'm violently opposed to ads on Twitter, so the fact that it got me to acknowledge any interest at all is pretty good.

    Curveball (Updating)
    A Rake by Starlight (Updating)
  3. Patrick Rochefort (Member)

    Posted 5 years ago

    I should add to that, for analysis, I decided to do my best to ensure this campaign would be served only to PC users instead of mobile users. As it is I'm not comfortable trying to target mobile users on Twitter, they seem the least likely to be reading webfiction.

    (But hey, I could be wrong.)

    From Winter's Ashes: A Detective with nothing left to lose, against a Necromancer with a world to gain.
  4. Jim Zoetewey (Moderator)

    Posted 5 years ago

    Do you use Google Analytics? If you do, check the stats. Personally, I sometimes find that as many as half my pageviews are from mobile devices (phones and tablets). Occasionally it's more than that.

    I"ve been watching stats for that since I realized that people could read my site on their phones. As a percentage, it's always heading up, and never down.

    EDIT: That said, while it sometimes gets up to 50%+ on a particular day, I just checked my total mobile stats and found that it's above a third overall. I haven't calculated the exact percentage.

    For me, it's very much worth checking how the site works on mobile because it's definitely getting used--sometimes for hundreds of pages.

    Related to that (and I wish I had a good citation), there are actually studies now that show that people with dyslexia have an easier time reading on their phones than they do actual books. It's something to do with having less wide pages, making it easier to keep track of where they are.

    Thus you might actually find that you have readers that couldn't read you otherwise.


You must log in to post.