A few things that struck me. Note I primarily do Facebook consulting so I'm pulling broad generalizations from there but I do run youtube ads as well.<p>1. Using predefined broad audience categories rarely works. Doing keyword targeting or specific video targeting usually works out better.<p>2. $10 cpm isn't high. I regularly spend $40+ on Facebook that can be 5x+ profitable ROI. Judging an ad buy only by cpm is short sighted at best.<p>3. 14k is a very small budget. I regularly spend 5-10k per day for folks. It takes volume and frequency to optimize. Even the amounts I spend as a freelancer makes me a small fish in regards to any serious agency and their clients.<p>4. It is far far too early to measure the effects of the campaign. CLTV (customer lifetime value) takes years to realize. Modern analytics tools are very useful to inform us on performance. They are not the end all be all and don't do an awesome job of multi touch attribution.<p>5. 1 ad isn't really a great test. For clients spending any real amount of budget I can pretty easily push out 300+ ad variations a month. To find something that gets you the kind of results you want it takes time, effort, and paying close attention to what's working and what isn't.<p>Overall it's an interesting case study but I caution folks to not simply take this as "youtube doesn't work" but rather what this person did on youtube doesn't work.
People inexperienced in online adversities often think that they can pay for conversions. That is not the case. You’re paying to get your message out. Very very few people will immediately act because of an ad. But if you hit them enough, they may later convert.<p>Tracking those late conversions is imprecise. This does not mean advertising is useless. But it is not generally a game for those with a tiny budget or a short time horizon.
Well the initial snap impression is that the ad isn't very good. But let's talk about how this could have been done better.<p>For starters, I would not pitch a real world get stuff in your mailbox service with animation. There are no photos of the product, the explanation of how it works and why it's not a scam is hand wavy. I don't care how it works I care about <i>what's in it for me,</i> the buyer.<p>There are no clear examples of what you'll get. There is no sense of <i>who</i> is behind this service and why you should feel connected and like you would trust them.<p>In addition the implementation tactics are suspect. You targeted hardcore gamers, but you're not selling games. That's almost a complete non-sequitur.<p>Your goal was to target people likely to purchase candy, or perhaps Japanese novelty items, via a mail subscription service. It doesn't appear that you made much effort if any to figure out what that audience looks like and how to reach it.<p>One of the reasons digital advertising can be so effective is that it allows you to try multiple different approaches at the same time. Nearly every successful PPC type campaign starts with multiple theories, different hypothesis about who the audience is and what they'll respond to.<p>You try a few ideas, see what works, track sales and intermediate metrics (clicks, engagement, time on site, add-to-cart behavior) and then continue to direct your budget to things that show signs of yielding positive ROI.<p>These things are to some extent always a roll of the dice, but there are definitely some things you can and should be doing here to improve your odds of success.
Hi Bemmu. I subscribe to TokyoTreat, and formerly subscribed to Japan Crate (and the 'umai crate' noodle box, until one of them got seized by customs because they shipped something with pork bits...)<p>I don't subscribe to CJ despite having seen it a million times here and almost definitely being the target audience (though this particular ad would have definitely repelled me) because your prices just aren't competitive. I mean, sure, the root price of $29 a month is, but when you consider the volume of candy you get, it's not even close. 7 items on your example image vs. the 14 I got in the February TokyoTreat (one of which was a heavy drink and another of which was a 30-piece share pack) just makes it clear.<p>Have you considered dropping the shipping to once a month like all the other boxes and making them bigger? I think that would help you compete more effectively, because you'd be paying less for shipping as a percentage of your subscription gross and the "unboxing" pictures of your boxes wouldn't look as sad.<p>$29 for 7 items is over $4 an item. You're going to have a hard time convincing people it's worth paying that if they look at any other subscription box.<p>From looking at the list of past boxes, I think you do a really good job hitting the right balance of "ooh, exclusive Japan thing" and balancing out the tastes of an average person. The candy you select is spot-on with what I considered some of the best stuff I got when I lived there.
The ad was terrible. It explained how Candy Japan works. I don't care how Candy Japan works. I want the why should I buy Candy Japan. It should have focused instead on the candies that you can get. Delicious, weird, unique candies only available through this service.
While the Candy Japan lost the revenue on YouTube ads, I wonder how those posts, that hit front HN and certain Reddit subreddits convert.<p>Would you make money back, if you would include all the signups you will get within next 96 hours? The "mistake" cost you real money, but on the other hand, it allowed you to generate a content on a very interesting topic, that surely will get you thousands of new visitors, right to your website.<p>Also, the animated video is... well, it is not good. You can easily get highest quality animated videos on Fiverr or PPH for way below $1K.<p>I would suggest you get featured on unboxing video of some semi-famous guy. I am sure, you will get much more conversions for your buck since those channels are watched by people willing to buy a product.
Honestly, I have no idea about ad campaigns. However, I've felt that the ad itself wasn't good enough.<p>And, when you say "a typical YouTube ad is 30 seconds" doesn't mean you have to make one so long, you could have considered it more as an upper limit instead of a goal. Your message was simple, and i've felt the ad was unnecessarily slow and long. Being more specific, you lost me exactly after the first 8 seconds.<p>On the other hand, the ad taught me that it's possible to buy japaneses candies easily and where i could get them. If someday I, or one of my friends, want japanese candies, I could simply point to your website. (And in fact, I already made a friend interested about it before reading your post).<p>Lastly, the ad itself is boring, maybe with some background music or even a voice more like "Hatsune Miku" could been a little more fun to see entirely.
Speaking as a marketer: getting positive ROI on ads is hard. That’s why people hire professional marketers to market things. I often find that non-marketers think they can do marketing, and then when it doesn’t work, they tend to blame the platform. I hire engineers to do engineering work, and likewise business owners should hire marketers to do marketing work.
As a marketer, videographer, and someone with a sweet tooth, it's not a great ad. Get to the point. Show the actual product. Tease me with things I can expect to receive. Maybe show reaction videos of people opening and trying the candy (and loving it).<p>Also, you'd probably get higher ROI sending $14k worth of Candy Japan to YouTubers in the hopes they'd open it on their vlog and mention the website.
Good post, some nice lessons there for someone entirely new to this.<p>Perhaps a missed opportunity was not advertising the domain name during the entire video. Often a user may not have the volume up and so perhaps providing a textual call to action in the first half continuously would help with conversions?
>But the cool part is that you can also track when someone views your ad, and then a bit later types in your web address to make the purchase without ever clicking the link.<p>>In other words you can track both people who click through an ad directly, or who see your ad and then visit your site a bit later<p>How does this work?
I really like the honesty about how he lost money buying youtube ads...feels like there's definitely a market for an self-serve "pay x amount a month" for "y results" app if you will.
An unfortunate part of experimenting with advertising is that it's expensive. Trying to find the right audience for the ad is very costly, to the point where I don't understand how advertisers can afford it.<p>I say this as a tinfoil-hat person: narrowing the audience is going to need a lot more info about the person. Particularly, their browsing history (i.e. remarketing). Knowing this would raise the conversion ratio immensely at the cost of privacy.
Very useful post thanks. I didn't like your ad though. It doesn't show me the product. Every food item ad I've seen: slow pan on the food product and then someone eating it - inserted somewhere in the sequence. (Given that they all do this, I would bet there are million dollar marketing budgets somewhere with data to support that this works.)
Instead of spending 3k on the animation, you could try the way gaming companies does it.<p>It looks to me that they grab the first 10 people off the street, pay them $10 to read a script (really bad awkward scripts as well!) and do 10 videos like that.<p>They are everywhere so it must be working and it must be cheap as well.
The author can't really be totally certain if they lost money though, since it's entirely possible the ad will make an effect on people long-term. Maybe they'll break even in a few months, who knows.
2 observations:<p>1) A video ad is potentially a poor environment for direct response signups in this case. People usually just want to watch the video they clicked on (unless maybe you're advertising on videos about Japanese candy — but in that case it might me more effective to sponsor videos by a popular video producer / channel (what SquareSpace does with Lessons From The Screenplay etc). At least with a sponsorship you can have a link in the description that doesn't disappear in 30 seconds.<p>2) Going a bit further, something like this is ideal for content marketing on Instagram (et al). I looked for an example of an account that posts novel Japanese candies. Low and behold the top two are both content marketing for similar services.
<a href="https://www.instagram.com/tokyotreat/?hl=en" rel="nofollow">https://www.instagram.com/tokyotreat/?hl=en</a>
<a href="https://www.instagram.com/japancandybox/?hl=en" rel="nofollow">https://www.instagram.com/japancandybox/?hl=en</a><p>$14k will get you a lot of pretty photos, posts, hashtags, follows, comments, etc. You continue to own your audience going forward and you end up with a neat little online publication / channel, which is a great jumping off point for other types of organic publicity (news quotes, etc).<p>Disclosure(?): I do content marketing / PR and formerly worked for a video platform on the product side.
While most of us happily discuss what Candy Japan's CEO did right and wrong with his Youtube ads, they miss one little thing:<p>Candy Japan, the master of upvoted stories on HN, burned $14K on Youtube to have another story for HN which brings attention, traffic and new subs. Still: The thread is full of super important infos, so all good.
Speaking to a few anime/manga fans I know Bemmu has to compete with giants like J-List who dominate this space. Then there are local "Kawaii companys" in each region, for example Tofu Cute in the UK.<p>Why doesn't the Candy Japan Website have prominent links to Twitter and other social media?<p>Have you thought about using Social media more? Let's look at Twitter. Bemmu has almost 6,000 twitter followers, I'm guessing the Candy Japan Twitter account is @candyjapanese with about 250 followers?<p>While rivals - <a href="http://www.japancandybox.com/" rel="nofollow">http://www.japancandybox.com/</a> has 23K follows and J-List has 155K followers.
Bemmu presented this very same topic in person at a previous HN Kansai: <a href="https://www.youtube.com/watch?v=hs30WPkRpsY" rel="nofollow">https://www.youtube.com/watch?v=hs30WPkRpsY</a>
As others have mentioned, the ad isn’t a good use case for this kind of blind marketing. There is simply no hook, evocation and it’s not emotional. It would work better as a landing page video to further explain the service to an already engaged potential customer, but not to new prospects targeted with route filters.<p>I think it could work a lot better as is if it was used for retargeting. i.e. people who have already been cookied and visited your site but haven’t yet made a purchase, or clicked through on some of your email marketing.
I wonder if there is "brand awareness" element that resulted from this ad campaign.<p>It's not immediately obvious - but it may be a prolonged subtle benefit resulted from it.
I don’t understand. You didn’t do any alternative testing? Why not?<p>Whenever I test a new channel I start by making variations to test different factors, so that I can learn what works and what doesn’t. I feel like a lot of people run ads but don’t quite get the concept of running a campaign.<p>My suggestion would be to try that next time... while you didn’t make money the real question your left with is what could you change to actually make money?<p>Loved be writeup! Thank you.
I only spent $75k in my life on ads but I had a x2 average ROI on it. The first thing is: you need at min. 30 different ad variations. Copy, animation, video, value proposition, pricing, landing page. Same with audiences. That's just the minimum. Therefore, standardized ones are obviously a no-no. With 1 variation, you've just - as you rightly wrote - burned the money.
I agree with some of the other commenters the Candy Japan is a product that appeals to the soul not our brain. I doubt there isn't a better ad that could have been made to sell Candy Japan more effectively at a lower cost.<p>This should not be a lesson in avoiding all ads, not all ads are made equal.
I'm not sorry to say that Candy Japan will never see any of my money for the simple fact that there's no way that I'll sign up for anything that expects a recurring charge. The whole "service" attitude is flawed in that this is not a service, software as a whole is not a "service" it's an end product. Good for everyone whose pulling money out of these schemes but not a single cent of it will come from me. And $29/mo * 12 = $348 per year. (forming assumptions about the contract i.e. predatory "cancel early" clause and similar) On candy? give me a break, I don't spend that on candy over 5 years even if you include other needless crap like soda and sports drinks I doubt I'd hit that number over 5 years.<p>I should also add that I'm the epitome of anti-advertisement. The more that I am forced to see (thinking of YT style of ads here) an ad, the more likely I am to NOT* use that product or service. In fact, the more likely I am to purposely avoid that product out of spite.
14k is nothing. Lookup how much a single billboard beside a highway costs.<p>I watch youtube all the time. I some ads, at least the first few seconds. I have never clicked on one deliberately and certainly would never be a convert, but i do know these companies exist. That is worth something.
Great read. I am in a similar place with a small business. We have a commercial and animation and are considering doing YouTube ads. It sounds like it might be worth doing if the service is profitable enough per user. Really appreciate this detailed writeup of his experience.
If I were going to spend 5 figures on an ad campaign I wouldn't <i>write the ad myself</i>, I'd hire a professional marketer. I think that's the biggest mistake: you only get a few seconds to capture someone's attention on YouTube.
I posted this [0] on you 2017 review post. Wish you would at least try it. For some reason I think it could work.<p>[0] <a href="https://news.ycombinator.com/item?id=16015617" rel="nofollow">https://news.ycombinator.com/item?id=16015617</a>
Since it’s a bidding system, someone’s getting good returns on these ads and I’d speculate it’s just a small handful of people who actually hit a sweet spot, with the rest filled by people doing temporary campaigns or testing out campaigns.
The number one rule of spending money on digital ads is to spend a small amount on many different ads before spending lots of money on one. This is doubly true today when a guy will make you an awesome video ad for $25 on Fiverr.
> In other words you can track both people who click through an ad directly, or who see your ad and then visit your site a bit later (this is known as a "view-through conversion"). It's magic.<p>It's not magic, just evil.
Since a lot of marketeers seem to be here. I have a small webshop, anyone willing to improve my monthly ( 400 € revenue) to more through ads on commission ( adwords only currently)?<p>Only handling dutch currently, niche - swimming
IMHO this is the right stuff if you want to sell japanese candies with a video:<p><a href="https://www.youtube.com/watch?v=g6FosltlFoo" rel="nofollow">https://www.youtube.com/watch?v=g6FosltlFoo</a>
I love Candy Japan and the articles are always fantastic. However, I can't help but wonder what about CJ is so captivating to the Hacker News audience - the posts seem to always end up on the front page.
><i>YouTube allows you to pay to show an ad before a video plays.</i><p>Is also the reason why I installed an adblocker.<p>I do not know if anybody still remembers the time when ads were not intrusive and did not try to hog all the CPU cycles.
A good approach might be to just produce videos and integrate them with the blog until you hit something that looks like a system and relatively viral. Then with that content go to the ad networks.
> I assumed to trigger a keyword match, the video title or description would have to match it. This turned out not to be the case.<p>Does Youtube extract the keywords from the speech in the videos?
could people with a love for Japanese candy do for me what Candy Japan couldn't, mainly suggest some great things to try? I've ordered a ton of different black licorice/salmiakki from Scandanavia and Germany (fazer salmiakki filled chocolate bar? oh my Lord yes) and need another adventure!
tldr: "In conclusion, did you break even? Nope. I lost money."
"<p>I too have lost money on adwords, and have a feeling it is more the rule than exception.
I'm interested if a free spot at the top of hacker news hot articles will end up being better in terms of advertisement than spending ~$14k on youtube ads.
I love the marketing industry. Everyone is overfishing in the same ocean. People's income don't increase much year to year but internet ad spending sees double digit growth!<p>My budget allows me to spend €500 on entertainment/leisure per month. You can bombard me with advertising all day but that's not going to change my finances.