Unit economics never works out for gifs. It's historically not worked out well for normal JPG image hosts. It will most definitely not work out for hosting files in a teriibly un-optimized video codec.
I think the whatsapp deal probably works out -- but how many of those will such companies bank against.
One hope could be that disk and network prices become low enough to sustain some small profit, but when that happens a dozen other hosts will pop up. A gif website isn't hard to host.
Despite being called "GIFs" they are actually delivered as videos (as is often the case today) - still, the money obviously is going to have to come from something else than pure hosting.
If I remember correctly, gifs are palette based and only 256 colors. They also don't have alpha transparency, instead having 1 transparent entry in the palette, so for a gif to smoothly blend into its background it has to be pre-rendered against that background.
In addition to being limited to max. 256 colour palettes and 1-bit alpha (encoded as a palette entry), they only support run-length encoding (Edit: NOT! it's actually LZW) for compression. The combination of RLE (Edit: NOT! It's actually LZW, same point somewhat valid) and dithering (used to overcome the palette colour limitation) results in typical compression ratios which are worse than nothing.
Personally I think a hybrid between a video codec (h.264, VP10, etc.) and a lossless transparent still codec (PNG) would be a good way to go
GIFs use LZW compression, which is the part that was patented. I think the main problem is that the dithering necessary to make 256 colors look good introduces a lot of high-frequency noise, making the compression less effective.
Oh boy, I don't know how I got the impression that it was RLE. Maybe I was mixing it up with older versions of Targa.
Yeah, the noise was more the point I was getting at. IIRC PNG also has the benefit of prefilters which allow it to (and maybe there are other things) compress rectangles as contiguous symbols rather than splitting them across lines, which makes obvious sense for 2D images.
IIUC APNGs compress the delta between frames. So 2 frames in a video have mostly similar pixels, and so will compress very well. And even on still images, PNGs have much better compression. From the same principle that nearby pixels usually have similar values.
Gifs store every frame as a separate image and use very crude compression. Also the very limited color palette as others mentioned.
Quote: "Imgur has always been profitable right from the very beginning and the only time I ever had to spend money on it was for the initial domain name because as soon as I released it, people liked it so much that they were donating and so Imgur survived for the first six months just purely on donations."
I mean now that they've created a social network and are redirecting all image links to their shitty webpage with the cat paw, ok, I believe it, but since day 1 when they were just hosting images? There's no way they were making money off that
I liked a lot of what Giphy was doing with TV show parsing and captioning, I feel like that could be a huge vector for sponcon, but this announcement about sponsored GIFs and ad sales sounds go generalized on its face that I'm not sure it will be fruitful.
Is BMW going to buy sponsored GIFs? Is McDonald's? Are major CPG companies going to buy sponsored GIFs? I think TV shows and major studios have a good use case for this, but even if the sales are constrained to suitable markets, and healthy, will those sales support a $600mm valuation, or even keep the lights on?
Ad sales work well when you control a platform with a lot of eyeball-share, and when you have strong profiling. I don't think Giphy will have much of either, and if they do, I think the they won't have all of those attributes available (i.e., giphy.com is hard to build facebook-level profiling on, but is easy to control ad units on, GIF keyboards/facebook/various integrations are high-eyeball, but someone else controls the profiling and you can't control the ad units etc).
I could see them being a clear buyer here. Some of the GIFs in that article are clear, in-your-face ads, but some of the others are a bit more subtle... yet still probably achieve Starbucks' ad goals.
As for Giphy reach... I use Giphy on Facebook Messenger, Discord, and probably a few other places that I'm unaware of. They hook into a LOT of places.
"hook into a LOT of places" - That optimism isn't well supported by acquisitions and IPOs in recent years. I maintain that the "pop-culture-image sharing" craze, which obviously includes gifs, hit its monetary-value peak when Yahoo bought Tumblr in May 2013 for $1.1B. It's been more than 4 YEARS since then, and what goin' on with Pinterest? I mean they are now 7 years into their existence and have neither IPO'd nor been acquired. Meanwhile, Imgur keeps chugging along--though downward, and since Reddit began hosting its own images, Imgur's decline has only accelerated, as even the Google Trends popularity graph reflects [0].
VIDEO is the future, and with AR, VR, and mixed reality on the rise (where the Almighty Facebook's Oculus is actually 5th place in sales [1]), images and gifs are looking about as exciting as innovations in the design of mobile flip phones. Snap, with Snapchat, went public earlier this year, but I don't think they ever could have done if they remained merely a photo-messaging app, rather than a photo-and-video messaging app, as well as a news-publishing app and now a TV app with its own proprietary programming.
Didn't we learn from Zynga that a business can never be truly successful when it completely depends on the participation of partners to have any successes. Facebook could decide tomorrow that they no longer want to give Giphy the power that they do, and completely destroy them no?
I use Giphy daily, multiple times a day, on Slack. Unless Slack is paying a hefty SaaS fee to Giphy for use of their platform, I don't know how else they'd monetize it.
I see McDonalds integrating Giphy into a larger ad campaign. Let's say they launch a campaign around happy moments with family. That gives them scope to includes moments from the TV or cinema spots as featured results when someone searches for a happy or family gif, which the user is more likely to do because they're aware of the campaign. That user then shares the campaign to friends on social media who may not be television watchers.
Or consider the next Minions movie. They just animate and render a bunch of Minions acting out the top Giphy search terms (subtly tagged 'Despicable Me 4'), and suddenly the movie is being organically promoted by thousands of users.
Some of the HBO ones from Silicon Valley had to be sponsored. If they weren't they're missing out for sure. Definitely, think tv shows and movies are the way to go here. A McDonalds gif just doesn't feel right. Almost all of my friends just share clips with captions from shows.
Good luck living up to that $600 million valuation with sponsored gifs before running out of money. Seems they raised 55 on 300 in February 2016 and then 72 on 600 in October 2016. Seems like a lot of money to be burning for a gif search engine.
I think there is a place for a company like theirs, at a much smaller valuation, running and making profits. But now that they took VC money on such a high value, they're bound to do all the classic tricks to burn through money keeping the illusion going ...
But if they only make $1/year profit on each of their 200M DAU, the valuation is easily justified. I'm sure FB, Google make much more than that and I haven't clicked on their ads for years.
FB did over a $5/year on a monthly active user in 2016. But they (and Google) are much better positioned at selling ads than a funny gifs engine ever could be.
These are still 200 million pairs of eyes every day on their product. VC funding has looked like a joke recently, but 200million daily users at $600M doesn't seem ridiculous, regardless of how silly/bad the product is.
But that's just users viewing content there, not interacting with the site at all. I'm sure imageshack had those numbers, easily. Until something better (read not monetized) came along to replace it for a while.
You might be right. I've never thought of giphy as just an image hosting platform though. It's more about the integrations with apps that is their edge. Something imageshack/imgur never really nailed.
But I think there are variations on this that work/make money.
`/giphy football` could return some relevant superbowl sponsored gif as one of the first N gifs. Kinda like how google gives you 3 "relevant" ads before showing you the actual search content. Start subtle, get more and more brash as time goes and people won't notice. I guess making it work without people leaving is the $600M question.
Yeah totally agree, raised too much money and won't be able to live up to the hype. Probably could have been bootstrapped into a modest company instead.
And for a really bad one. It's more of a gif aggregator that can give you gifs of approximate themes. But I very rarely find a specific gif I want and I know it exists.
Use RightGif, much better IMHO. It can find really obscure GIFs with ease.
Giphy's Slack integration is a pain point, just pasting in the first result (or just a slot machine type system) is not the best UI. They do better on other platforms, but RightGif tends to get it right the first time.
That may make the task more difficult, but still, even when I search gifs from very well know series like the Simpsons and I get a bunch of random Simpsons gifs (therefore I assume they got rights for short Simpsons clips) but none is remotely similar to the one I'm looking for. Then I go to google images and there it is among the first results.
Edit: It may have improved a bit lately, but it would be much, much better if they did OCR of text in gifs, and use the context of where the gifs are found (esp. page title) for additional search terms.
Yes, currently I know it and use it. It was just an example on top of my head. Other series don't have an equivalent. And of course there's a lot of known gifs that don't come straight from TV. RightGif mentioned in another comment seems to work better for that purpose.
How can giphy with its current functionality possibility cost that much to run? I feel like it should be able to run for over a century with the amount that it has in the bank right now. Bandwidth is cheap and it's not exactly compute intensive or labor intensive.
I could distribute that thing across a global cluster of VPSes with cheap bandwidth (Vultr, Digital Ocean, OVH) and host it for peanuts. Given that it's drowning in VC money my guess is that it's all at AWS, not using S3, and over-engineered to Juicero levels. They're probably using Kubernetes, a dozen different database and message bus protocols, and a whole sprawling Rube Goldberg machine of a microservices architecture to... host gifs.
... and yeah that nose candy valuation is ludicrous. There is no way.
200M users, 20 MB per user daily (optimistically) gives 4 petabytes per day. I'd say the average user transfer might be way higher in reality so we're talking here dual digit petabytes on daily basis. How much will it cost you on Vultr, Digital Ocean or OVH? I'm asking out of curiosity because I'm not much into hosting, but I think it might be quite expensive after all? ;)
You really think 200M people are transferring 20MB/day of gifs?
If your numbers really are accurate then it underscores the insanity of the valuation. If the costs are intrinsically high then you'd need a high value advertising niche to make this profitable.
I'm almost certain giphy is already testing sponsored gifs, or someone is dark patterning giphy's search engine. Every once in a while in slack, /giphy would return a gif that was 100% an ad and 100% not relevant to the parameters passed.
> Especially if Gif searches sometimes have intent.
...they do? I'm not sure I've ever seen a Gif posted that was not a pithy reaction to a conversation. Seems kind of hard to come up with meaningful context from "spongebob nope".
I think that number one, your assumption about search volume is very hard to ascertain: do they have 200mm DAU? MAU? Total users? What's the total search volume per day? Is it ~450mm? Are all of those searches monetizable? Is the revenue per search on par with Google's?
I think there are huge gaps to fill in before we can understand what the value of monetized search on Giphy is vis a vis Google.
I think a lot of people here are pessimistic about Giphy just because they took on huge VC funding, but if Giphy can maintain and grow its user base and be able to monetize and earn a profit of .5 dollars per user per yr, then it would have earned 100M in profit. I know it will be hard but thats entrepreneurship, its never easy
Giphy is such a useless product, I have trouble imagining them being able to monetize for more that 0.10/user/year.
Not only is it a timewaster application (why would a business pay for employees to use it), but it also does a poor job of search - its main selling point. Baffling that anyone believes it's worth $600M.
I think they could only solve their search problem by hiring an army of minimum wage gif classifiers to curate the attachment tags and labels to each gif.
That would make their search usable, but still not sure what the market for gifs is. I don't see any good way to compel their end users to pay for the service, and ads will fundamentally degrade the utility of the service even further.
They've essentially built a poorly-functioning toy and are getting investors as if they are a B2B SaaS service. What is their value-add? Is there any economic utility in being able to quickly find a gif?
> Is there any economic utility in being able to quickly find a gif?
Giphy gets used all the time in my company's Slack channels..
It's a morale thing I think (at least a bit).. People enjoy putting a bit of fun into the work chat, in a way that's acceptable (in our office culture - don't want to speak for everyone) and low-effort..
If Giphy was suddenly 5-10$/month (like RightGif is for example), I think my company would gladly pay that to continue to let our employees use it during their work conversations.
Fair enough. In my company's experience giphy only attaches a relevant gif around 25% of the time, the rest of the time the gif is an odd non-sequitur. Relevant and amusing gifs are probably closer to 5-10%.
We used it a bunch when we first moved to slack, but now we probably use it less than once per day across the whole team. We'd turn it off without hesitation if we had to pay for it.
This is a mere personal anecdote so it doesn't mean much at large, but I'm pessimistic about the advent of sponsored {thing} not necessarily with Giphy. To me this is just another %thing% that advertisers have leached a tentacle into for the sake of getting me to buy something.
I get it, it's the game we play now with content that is ostensibly or otherwise "free", pandora's box is open but the internet has become such a frustrating place for me personally every time I see "Sponsored tweet" or "sponsored post" or "sponsored ________" that tries to blend itself into the rank and file of non-sponsored content I would otherwise enjoy.
The question is, how can the non-sponsored content you enjoy continue to exist if the company that hosts it doesn't make money?
After the VC money runs out, content hosting service need to pay the bills somehow. I can only think of a few ways this can work: you pay for the content you like via a subscription or per unit of content, you look at ads, or you look at sponsored tweets/gifs/whatever. I suspect that most people would be unwilling to pay subscription fees, so ads and/or sponsored content are probably the way to go for most companies.
Agreed and well met. At least for me, if the content generated is truly engaging and substantive enough, I gladly would pay a subscription fee. This may not work arguably for a gif site, but your point is well taken-because it's completely true and we see it daily.
I suppose to unpack my position a bit better than I have so far, if the sponsored content is truly relevant and engaging, I'd have far less things to complain about on this topic.
As it is, the majority of "sponsored content" I've seen hasn't really amounted to much that seems to fit within the ecosystem it exists in; it oft seems out of place and irrelevant.
But again, your point is well taken and I'm sure smarter people than I will continue working on this question and find ways to continue delivering content freeloaders like me love to consume :P
I'm hoping it will bring better search functionality, but the search will probably still suck plus ad gifs pasted throughout. Their search functionality has autocomplete as if it might know what you're looking for. The results are typically mediocre though. So here's hoping that they use the ad money wisely to improve the search algo.
I wonder if there's a potential issue with trademarks there. If "Netflix and chill" becomes popular enough, does Netflix end up just being a generic word for watching online TV?
Depends on what you mean by it.... people do definitely use it to imply a sedentary night watching tv, although others use it ironically to indicate a booty call.
I didn't know this wasn't already in prod. On numerous occasions, I've seen less-relevant gifs show up with an album link to a brand or product. It was what I expected the justification to be for the poor accuracy of gif search.
Do Giphy actually host Gifs anymore? I noticed how they make it very hard to actually access the image version, and even this image is in webp, so you can't embed it. It's all iframes.
This is not saying they're bad, I totally can relate with their need to cut costs because just hosting these gifs for free and letting anyone <img src> them is not a profitable decision, but then again, it's not really Gif anymore.
So even in this case the "sponsored gif" would probably be some sort of a video instead of an actual gif.
Nowadays whenever I want to search for a gif I want to use for a blog post or website, I have given up looking for gifs on giphy and just go to google search instead. The "gifs" on giphy are unusable for my purpose.
I see your point but don't think it's an adequate analogy. Xerox still means photocopy and Hoover still means vacuum cleaning. Gif is something that used to exist for decades, not something giphy invented.
Gif is far more useful as a media type rather than file format; and more than that, i don't think most people realize or care that they're actually watching some kind of video-sans-audio.
You know exactly how much each GIF served costs you.
Charge that to a balance I keep on file on my account, and add autopay with a monthly spend limit. Let me indicate whether my GIF links should become sponsored or should simply return 1x1 transparent when my balance for the month runs out.
I don't want a "premium monthly fee" account, I want you to charge me S3 delivery fees + 15% markup for the privilege of dealing with S3 for me.
What if someone hotlinks your gif? Are you planning on using some kind of domain whitelist? How would that work? I like the idea, I just think enforcement is difficult.
Half the fun of using giphy through Slack was the risk of getting an awkwardly bad gif. Once there was a preview, folks in my Slacks stopped using it.
I use the giphy website a couple times a week, but the search is terrible. It's my go-to choice purely because it's convenient, but there's no way I would ever pay for the service. If other users feel the way I do, sponsored gifs may be their only hope.
Seems like giphy is counting some things they shouldn't. If a giphy gets embedded on the page, maybe they count every single impression of that gif as a daily user?
I'd be shocked if anywhere near 200M people actually had accounts on the giphy website.
> If a giphy gets embedded on the page, maybe they count every single impression of that gif as a daily user?
This wouldn't surprise me in the least. But, if they can turn a reasonable number of those impressions into sponsored impressions, it's kind of beside the point.
Although doing the numbers, 5% paid impressions on 200M daily impressions at $5 CPM is only $50K daily or $18.25M yearly...
The moment they replace a gif image with ads it will kill all that goodwill. I do think selling the gifs would work though. People (1) are willing to pay a lot for attention/be cool/social/funny.
The strategy will be to purge the Internet from gif images containing content from their partners, so that you can only find gif images from their site.
1) I'm glad HN doesn't allow gif images in the comments. If you are reading this you are probably not in the target group that might be willing to spend money on gif images.
We've certainly come a long way from when websites with gifs were annoying in the early web days to where there's a company with a GIF raison d'etre and a $600M price tag on it.
I originally was picturing like a normal gif but with an opening, intro ad. I think that's incorrect.
I think ad-sponsored GIFs will be more popular and more legal then any other type of gif. I envision a future where we are all sending ad-company created gifs to each other without even noticing that they're created by ad companies.
GIFs are so easy to make and so cheap, and they're used everywhere.
I wonder how this will interact with them being sourced by Facebook's gif functionality, which has already broken a few times in the past month or 2, requiring you to physically go to Giphy's site to get a working link to the gif that you can paste manually in a Facebook comment
Part of it also is that Giphy has licensing agreements with media rights owners to create gifs based on their content . . . they're not just indexing existing gifs.
I think it's a clear case of using "can't" to say "doesn't bother to because it's not worth it". In fact, I wouldn't be surprised that Google had great indexing of GIF already, but doesn't fully expose it in its search engine.
I think the whatsapp deal probably works out -- but how many of those will such companies bank against.
One hope could be that disk and network prices become low enough to sustain some small profit, but when that happens a dozen other hosts will pop up. A gif website isn't hard to host.