ne of the pioneers of the early internet in the 90s, Yahoo was the widest read news and media website with over 7 billion views per month, the sixth-most-visited website globally. It famously refused to buy Google for $1m in 1998, and again for $5b in 2002, as well as turning down a $40b acquisition offer by Microsoft in 2008 before ultimately being sold to Verizon for only $4b in 2016. You’d be forgiven for assuming that Yahoo doesn’t advertise on Facebook, as did I, until I saw a tweet by Josua Fagerholm indicating they had some sort of advertising arbitrage play going on — which is even more unexpected given that Microsoft manages Yahoo’s search ads. Inspired by this tweet I decided to look further into their ads strategy to see what’s up.
If you’d like to download all of the ads featured in this post, we’ve made them available in this google drive folder, as part of our Facebook ads library resource. All ad examples come from the Facebook ads library tool.
What did we learn?
Yahoo’s advertising account is the weirdest one I’ve reviewed. What do you advertise when you could advertise anything? The answer is everything. Over 500 products by my estimation, and all built manually, with typos and emojis to stock photos and custom made videos. It was really fascinating to scroll through these examples, and I’m still not sure if this is a genius way to target a more valuable (gullible) audience, or just plain bad marketing. You decide.
If you’re interested in Facebook ads (if you’re reading this article, I presume you are!) then you might want to take a look at Spotlight. It uses machine learning to see what’s IN your images and text, then automatically show you what’s driving performance.
Diamonds In The Rough
This is the same ad from the original tweet that inspired me to look at Yahoo. It’s not particularly interesting as an ad, I mean I’m not sure that this is truly the best way to sell diamonds. The diamond emojis are one thing, the stock photos another, but worst of all the copy is uninspired and inauthentic. Just advertising a generic sale without details or even a list of brands makes this look almost spammy.
The fact that when someone clicks on this ad they land on Yahoo search results rather than a specific sale… is just weird. Like I’m sure the traffic from this ad will have a high bounce rate, because who expects to land on a generic search engine page? This probably works economically however — they’re arbitraging the lower cost per click on Facebook ads where users aren’t yet searching (i.e. have ‘low intent’), and turning that into high value search advertising traffic (‘high intent’).
Looking at Google Keyword Planner we notice something quite clever. The cost per click for ‘diamond necklace’ is 1/5th of the price of ‘diamond necklace womens’. This leads me to believe that the keywords they’re advertising here are from a list of the highest value search terms for Yahoo. It’s too specific not to be, and that strategy makes sense.
To the credit of whomever set this up, even if these aren’t the best possible ads in the world, they are following several best practices. For example they’re testing many different images to see what works best, and to give Facebook’s algorithm something to work with.
These are all stock images by the look of it, but then what can Yahoo do other than use stock images? They don’t own the products they’re advertising, and probably don’t have licenses to advertise specific brands. Fair play to them, if these ads are working then they’re working, and we shouldn’t criticise them for being un-imaginative.
One thing I would have liked to see is more copy testing. The original copy is pretty poor and it doesn’t change—they’re only cycling through different images rather than trying anything new with messaging. That’s a shame because there is a lot you could do with copy here, especially to try and prime the user for the Yahoo search experience.
Winter Is Coming
Diamonds are far from the only thing Yahoo are advertising. They’re promoting an enormous range of searches—over 9,000 ads with ~18 image variations per category, so approximately 500 categories. Let’s take a look at the latest one to be published in September, Men’s Winter Coats. You’ll start to see some similarities with the Diamond ads.
The ads are still using emojis and stock imagery, and driving to a search page for that category as an arbitrage on cost per click. Weirdly the link description at the bottom talks about DICK’s Sporting Goods, yet when you click through to the page it goes to the generic search results. I didn’t spot this on any other ads, so potentially it’s a mistake.
It’s typical that paid display or social ads convert to customers at a lower rate than search customers who have higher intent, and so while Yahoo makes money in the short term, I wonder if the lower average quality on some of these searches might depress the value of Yahoo traffic, and therefore cause advertisers to bid down on their CPCs, mitigating the arbitrage over time.
The low quality of these images is something else to consider—maybe these crappy stock photos capture more gullible people, who are more likely to buy and not be put off from this experience. It reminds me of Nigerian scammers who purposefully use misspellings so that only gullible people respond. Intentional or not, if that is the case, then improving the picture quality counterintuitively could actually hurt performance of this campaign.
Same goes for the copy—the Game of Thrones reference is a little off, don’t you think? Of course everyone knows ‘Winter is Coming’, but they don’t use it cleverly, in fact they follow it up with “It’s time to get ready!”. I don’t know if I’m reading into this too much but this sounds like something an older relative might say because they heard other people say it.
Targeting Older Demographics
Reviewing these ads got me thinking about the demographic of users of Yahoo, and something someone told me early in my digital career — “Yahoo traffic is more valuable because it’s only older people (who have more disposable income) that didn’t upgrade to Gmail”. If that’s true then potentially part of the strategy here is to target that older demographic. I don’t know if the poor image quality and use of copy is part of that, like the Nigerian scammers, or just bad execution, we’ll probably never know. But it is interesting that a lot of the products seem to be targeted to older people.
Maybe the search terms for Oxygen Concentrators are particularly profitable, but it seems like a weird thing to advertise. The picture for this was especially poor with bad cropping, but there were 17 other variations of Oxygen Concentrators in various guises.
Cremation, ok now this is getting strange. I can’t imagine advertising cremation urns on Facebook like this, but here we have it. Again I’m sure cremation urns are a particularly profitable category, given that people will pay anything when they’re bereaved, and these suppliers only have a handful of times in someone's life where their products are relevant. But also, is this something that actually works on Facebook? Surely the timing has to be wrong almost always… unless you’re targeting an older demographic with a number of their peer group passing. Too much of a coincidence.
Ok so this one was kind of funny, and hard to notice. If you look closely the headline of the text says “Best Jeans for Older Women”. Now this doesn’t make a whole lot of sense, the images in this ad are young girls. Plus when you click through the search page doesn’t say ‘older’ on it. So to me, this is a strange coincidence, or at least some evidence there is a purposeful targeting of older demographics (even if the image selection is a little off).
What did I say about the Nigerian scammers? They use spelling mistakes and low quality assets to only catch the ‘gullible’ people they can make the most money off… Ok maybe this is just a genuine typo, but it does make me wonder if this isn’t some purposeful strategy.
Travel to Aruba
Another category where the the older demographic over-indexes—travel. I know this because I used to work at Travelzoo where the older generation was key. They had budget and plenty of free time to fill. So if my hypothesis is correct about purposefully targeting older people, then this could fit the profile.
If you hadn’t worked in the travel industry you might not know that flight searches are actually pretty unprofitable relative to hotels (by a factor of 5 last time I checked) because airlines are notoriously low margin endeavors that frequently go bust. So if the strategy was just to target profitable keywords, would they really choose flights?
The ads in this category again fit the same profile of the others—stock images, poor adopy, emojis, generic adcopy. Of course this could all be down to incompetence: we have to observe Hanlon’s Razor “Never attribute to malice that which is adequately explained by stupidity”, but it did have me thinking. Especially when I noticed that when they did target hotel packages and not just flights, it was on specific locations like Aruba, a famous cruise port, on luxury all inclusive resorts, which are rarely booked by younger people.
One other working hypothesis could be that these ads are optimized to lifetime value. Even if they might not generate much revenue on the first click, the type of people and terms they’re matching here might lead to people who actually switch to Yahoo. That can be worth a whole lot more over someone’s lifetime. So it would make sense that the categories would come out as kind of random, and the ads that worked would be hitting people who are kind of gullible.
Random, Weird and Wonderful
Of course I could be wrong about this. The campaign could have just been set up by someone who downloaded a list of the most profitable keywords, didn’t know much about writing copy, and just wanted to brute force put 18 images x 500 categories live.
Rugs is one thing, I guess it’s a pretty solid category with a high average order value. I just find the copy on this one really bizarre. I had no idea hand-knotted rugs were a thing or that rugs had a ‘season’.
This one certainly grabs attention, but really makes me question my sanity. Paper towels? Really? It’s possible to profitably advertise paper towels Yahoo? When you only make a small cut of Bing’s revenues (who serve Yahoo’s results). Who in turn are only getting a small cut of the cost of paper towels. Which cost less than a penny each on Amazon?
The thing I like here is that they even made a video for the paper towels ads. They did this in a few other categories too, but it really made me chuckle to imagine someone pulling this together. Imagine the brief! One smart thing I did notice was that they used Promo.com for the video (I recognized the fonts / design of the overlay). I’ve used them for pretty size-able clients and you can do much better than this ad, I highly recommend their service!
I guess this is another in the category of ‘high value keywords’—I imagine this is a phrase that pretty much exclusively gets searched by CEOs. Though now I’m wondering how many CEOs they actually convinced with this ad.
Sure, quit your job, and then search on Yahoo to find your next one. You have to imagine the desperate job seeker thinking “great, some new job service, this could help” and then they click through and see it’s just a normal search results page, slightly worse than Google. This user experience is just cruel.
Luxury cars. Sure, why not. If Ling’s cars can sell cars with their website, why can’t Yahoo send people looking to lease 2020 luxury cars to their search results page? Especially when those cars have futuristic interiors and features.
Stickers? I am starting to lose my grip on sanity at this point. I wouldn’t imagine that anyone could profitably advertise custom stickers. Nevermind Yahoo advertising to get a cut of Bing-served ads which will is limited to the ads budget of the custom sticker tycoons. I don’t know maybe this is a higher margin business than I give credit.
Truck bed covers… now this one actually makes sense. This is probably a high margin item, that’s relatively niche, so I bet it has relatively low search volume, and this could be a way to juice it, to get people to consider a truck cover when they weren’t yet thinking of searching for one. That’s the key to this strategy—if Google has a monopoly on search, one way to beat them is to get people before they search in the first place.
Again this makes sense. Restaurant owners are probably big spenders, always Googling searching Yahoo’ing for new plates or other big ticket bulk items. I wish the ad was a little more inviting but honestly this is what these showrooms look like, so this will be a familiar scene for anyone making this kind of purchase.
Oh come on, you have to be kidding me. What even is that. How is there a USB with little tiny files on it? Or is this a giant USB with normal sized files. No that’s ridiculous. Is it? After just browsing through a few hundred examples out of thousands, I can’t imagine what it was like to build these Facebook campaigns. I do wonder after looking at all of this why they didn’t use dynamic product ads. They could have built a feed of millions of products, not just 500, and automatically generate the ads based on templates. It would have saved them hundreds of hours, and Facebook would automatically optimize for them. I commend their effort to do all of this manually (including the video about paper towels!) but there is an easier way!
If you’d like to download all of the ads in this post, they’re located in this google drive folder for this blog post, in our Facebook ads library, categorized by section. All ad examples come from the Facebook ads library tool.
Now that you’ve finished reading, I’m sure you’ll be inspired to run more tests of your own. When you’re testing at least 3-5 ads per week, you need a tool like Spotlight to help you get insights into the plus or minus factors driving creative performance.