Hacker News new | comments | ask | show | jobs | submit login
We Spent $1,138 on Facebook in 6 Days: Numbers and Lessons Learned (signnow.com)
125 points by shanecox1 on June 23, 2011 | hide | past | web | favorite | 67 comments

Am I missing something here?

The author seems pleased with the results, but as an outsider who isn't familiar with his company I see an ad campaign that cost over $1,000, achieved clickthroughs well under 0.1% despite targeting, and does not appear to have resulted in any measurable benefit at all. There are some comments at the end of the article about how Facebook is great for brand awareness, but there doesn't seem to be much evidence of that in the numbers.

If I were looking at ways to spend advertising budget for a new company (oh, wait, I am looking for...) then this seems like a pretty compelling argument for not wasting money on Facebook ads.

i thought the same thing. i spent a confused minute trying to figure out which units the 20/40 on the google analytics graph represented, given $1100+ in spending.

$250/day and on no day did facebook generate more than 40 hits. i would not consider that a success.

if facebook ads are good for "brand awareness" would it make more sense to use CPM ads with some sort of logo and/or memorable copy? they seem to be a bit cheaper.

Crap, that's terrible performance. We're probably spending $5.00 a day on Google adverts and pulling in about that many clickthroughs.

(checking), yup, we had 83 clickthroughs today and a total cost of $4.09.

CTR is at about 1.18%

(quickly scratches FB off of list of advertising vehicles)

FB and Google CTR is like comparing apples and unicorns :)

Typically with FB, you are doing well if you are finding anything near 0.05% CTR. Remember, these are content ads and not paid search.

Paid search you should be hitting 5%+ CTR on your top search keywords (in position 1-2, around 3% for #3 and 2% or less for anything below). On FB, getting an ad with a consistent 0.077% CTR means they are on to something and should keep going.

Don't knock FB off your list of advertising venues though. Each one is different and has a completely different dynamic, approach to users and value proposition. With paid search you bid on user intent through keywords - what they are looking for right now. On Facebook you are bidding on people and trying to get their attention, a big difference :)

With paid search you bid on user intent through keywords - what they are looking for right now. On Facebook you are bidding on people and trying to get their attention, a big difference :)

That's a very important distinction I hadn't really considered. Their method for "targeting" is entirely different.

I wonder which is more effective?

Well, each has a particular use and place. Typically, if you're selling widget software, you should start first with paid search by bidding on various widget related exact-match keywords. Go for people that are searching right now for your kind of product or service.

Facebook comes in to play when you know the sort of person you want. Their age, interests, location, education, etc. You can bid on them there and appeal to their situation, needs and interests directly. If you are selling XYZ service for race car drivers, you can appeal to people that love racing cars with an article and lead form telling them how XYZ service is highly essential for their profession.

Different approaches but still profitable. We spend about 40% of our monthly budget on search and 60% on Facebook. FB is very good when trying to generate leads that you eventually turn into sales. Paid search is very good when you want someone to take out their wallet and buy right the hell now, but you also will pay far more for that person.

It isn't clear to me that the users clicking the most popular version of their ad were not more interested in beer than in their product.

You got it exactly right. You can see how my numbers added up between Google and Facebook (http://blog.fairsoftware.net/2011/04/01/sell-your-google-sto...). Facebook allows you to attract people who don't know they could use your product (in my case, an entrepreneur conference - when is the last time you googled for a founder conference?). So yes, conversion rates are lower.

I call it discovery vs. intent. Google allows you to capture people's intent. Facebook allows you to let your target demographic discover new products.

Which is the Apple and which is the Unicorn?

They switch places depending on AdWords QS and FB's approval policies :D

The average click-through rate for Facebook ads was 0.063% in 2009 and 0.051% in 2010. Click-through rates used to be about 3% in the 90s, but have been declining since. Current industry standard is 0.1%. I heard that facebook is good for awareness, just to let low- or no-paying people know you exist, but click-through rates are pitiful, indeed.

CTR has declined since the early days but profitability has really increased. Check out some of the ad management tools if you're having trouble - they can really help you get into the data and find results :)

I can really see it going either way. I interpret this to show Facebook Ads are not a good way to get instant ROI. But there is a lot of value in having a large number of Likes because you can market to them each time you release a new product... in SignNow's case, they have 700 interested ears to begin each new adoption cycle now.

Hard to quantify the value of that. My opinion is that it's worth a lot for the early ears, but not if you are thinking short term only.

Facebook calculates an "edge rank" for each piece of content based on metrics such as likes, comments, post time, content type, and so on. Edge rank is used, among other things, to determine the likelihood of a given piece of content appearing in your feed.

Only a fraction of a page's fans will see each update from that page. By some estimates, that number is only 5-10% of a page's fans. So they've earned 700 new fans, but that doesn't mean 700 impressions each time they post.

Fair enough: if getting someone to Like you on Facebook turns out to be one step in a funnel that does convert a significant number of people in the long run, then yes, that has a measurable value.

I think the sceptic in me says that if you didn't get a worthwhile hit from a targeted ad spend, and if the Likers are only a subset of that target market, many of whom didn't even leave Facebook to visit your site, then it's unlikely to give a worthwhile conversion rate down the line either. Maybe the sceptic in me is wrong on this one, but only time will tell...

They mentioned the campaign stopped on the last day of the conference.

Why not let it run for an extra week or three? I don't think demoing new software you learned about during a conference is going to really happen by the last day of said conference.

Also if you did not make an impression on attendees enough to merit an @todo item somewhere, Facebook ads seem kind of futile.

mscarborough- This was more of an experiment for us, and we didn't want to burn funds indefinitely without understanding the consequence. Also, our best performing ad was a half joke about drinking beer at the conference, although we could have probably come up with some witty "our founders are hung over" type ad.

I don't understand why CTR should matter (except as feedback for whether their ads are performing or not) if they're paying per click (CPC)?

CTR matters for a couple of important reasons - profits & relevance.

Google, Facebook, Bing, etc use it directly to make more money on the same number of ad impressions.

Let's suppose there are 2 advertisers -

Andy who bids $1/click has 1% CTR & Bob who bids $0.60/click has 2% CTR.

When each of the ads are shown to users, Andy's will receive 1 click whereas Bob's will have twice as many clicks (every 100 times the ads are shown). The network (Google, FB, Bing, etc) will make $1.2 on Bob and $1 on Andy. That's some 20% profit right there without any special effort on the network's part while showing ads that interest more to the users. You only click something that interests you, right?

As you can see Bob, who bids a lot lower makes 20% more profit to the network for the same amount of ad impressions.

The second reason is relevance. One of the easiest ways to gauge it is CTR. Google, for one, places such a high importance on it that it has baked CTR right into determining how much you really pay to get that visitor in the form on Quality Scores. On Google if you increase your QS your ads start costing less and appear in better positions for the same (or lower) cost.

This is simple economics at play. If you see on the individual ad performance screenshot you can see that the ad that received 0.077% CTR has highest - - reach - social reach - connections - clicks and most important the least cost-per-click. Who wouldn't like to get all the benefits for a lower cost?

I understand that the CTR makes a difference to the network, but I am pointing out that a low CTR doesn't necessarily mean that the advertiser is worse off. It means two things: it means the advertiser's ad is terrible, or the targeting is suboptimal (relative to an ad with higher CTR), and that you are actually getting more impressions per dollar spent.

All else equal, you actually want to pay LESS per impression, given the same number of clicks per dollar (your CPC). Of course, in any well-designed dynamic system, the prices of the auction should reflect the different probabilities of being clicked.

I don't exactly want to hijack the thread, but let me present an idea of how one could burn $1,000 if one wanted to do it for promotion.

Step 1: Install Wordpress. Good, already done.

Step 2: Create a big ol' list of keywords which you think are relevant to your business. For example, [sign contract online], [sign document online], [sign contract with iPhone], [sign contract with Blackberry], etc etc. This doesn't cost anything more than your time.

Step 3: Design a template which, given a particular keyword, includes a few hundred words of content about the keyword and also acts as a landing page for your service. I'm assuming this is effectively free for most of us on HN. The last time I did it it required copy/pasting one file in Wordpress and then five minutes of hacking the PHP to do what I wanted it to do.

Step 4: Write between 50 and 100 of these, depending on how you get them done. This is what costs the $1,000, since you're probably going to zone out after doing this five times, write up the process, and then hire the rest out to freelancers.

Predicted results: on average, depending on query volumes, you're going to pick up a handful of hits for each page every month for life. So instead of having 1,400 clicks once, you'll get somewhere on the order of 400 clicks a month for forever. And instead of them being someone who was just gawking at a pretty girl in Chemistry class or poking their sheep while farming virtual cabbage, it will be someone who is right now trying to solve a problem which is costing their business money.

And if you have $2,000 to spend? It scales right on up.

And if you're savvy about how you do this? Suffice it to say that the MVP of this project gets you a wee little asset to build a business on, and implementing this strategy in a really effective way can basically carry a business on it's back. This is not limited to bingo card publishing empires, trust me.

I know you've answered this before, but where is your go-to place for quality writing? So many of the places that I've looked at seem Very low content.

The last thing I'd want to do is create a bunch of content that would make me look bad in the process. If I'm putting it online behind my company's name, it should be (hopefully) better writing than I could have done myself.

You don't have to be Nicholas Kristoff to write about signing contracts online, assuming you have a product which signs contacts online. Everyone on HN knows someone whose job description is Write Stuff About Business. Many of them, particularly young'uns, are underemployed in this economy. Find them, fix their problem.

I mean, you could walk straight into the English Department at any university and offer them jobs for students. I guarantee you will get takers. Potential problem: the takers will be English students.

I've tested Patrick's scalable content practices and have had much higher return compared to both AdWords and Facebook advertisements. Best part is you spend the time/money up front and it keeps working forever.

FWIW - In comparing these three methods (Content, AdWords, Facebook) based on their impact on the bottom line ($$$ -> that's what counts, right?) Facebook by far was the worst performer.

The only drawback to this is that it takes patience - results are not immediate.

Good advice overall, SEO is something to not be forgotten either.

I think you hit a very important aspect of this: each marketing channel (paid search, Facebook, Mobile, SEO, banners, etc.) is a tool in a toolbox. Go for what works and earns you revenue.

Paid ads yield dividends very quickly and are easily quantifiable. You can figure out what is working, weed out the bad stuff and then capitalize on opportunities. That said, SEO is also really important. While I am less inclined to go with the sparticle approach that many do, I would certainly say that writing up good information about your product in a way that relates to customers is extremely important.

The other part about your SEO content is that it is relevant to your paid ad campaigns. If you have a product that would work well for software developers and will save them money, then by all means write a good article about how awesome it is for them. Make the page a good starting lander for paid ads as well then get it to rank on SEO and drive other paid traffic to it.

In the end, each tool in your toolbox is there to make you money. Find what generates the highest ROI in the period of time you need to make it in. Focus on that and then circle back to each one down the list in turn.

Don't ever leave money on the table - your competitors won't make that mistake.

I have a feeling that this is the type of product that potential users don't think they need or don't know exists, which would invert the marketing method from finding users with intent to finding everybody and then converting them with your story

So I would have spent the $1000 on formulating a good story and reaching bloggers and journalists (which is what they did, or aimed at, in a way)

(btw what you are advocating here is an excellent idea, I just think it might not work in this case especially as they were looking to run a campaign in parallel with a conference and piggy-back on the conference name)

Edit: I did a quick check on traffic estimator, and not only is keyword search volume low, but it is very competitive. $5 per click in adwords, and the competition looks tough in the main search results:


Edit 2: a good idea would have been to buy keyword searches against 'techcrunch' and 'techcrunch disrupt'

This research should have been done as due diligence in the lightbulb->launch phase. If nobody is searching for the problem your product solves it's likely not going to be successful at scale.

I know you use this strategy with BCC. Did Panda make it less effective?

Not materially for my scale and strategy. I would suggest not doing it with 100k pages of spun content. If you don't know what that means, Panda is a non-event for you.

This is the author Chris, Great idea. We have actually been trying to do a similar thing with our "Ready to Sign" documents, but I like the idea. Thanks for the feedback!

Just a quick question, patio - do you make a separate domain for each keyword site, or just put them as pages in a larger domain?

I can't speak for Patrick but I do keyword based landing pages on the main domain where the product is for sale so you can pass page rank from keyword pages to your product pages or TLD.

This is also less work because you can create these optimized topical landing pages programmatically and just add some unique keyword rich content to them.

Also if you are adding domains regularly just for content you will have to wait a decent amount of time to let the domain age and you are always starting over at PR0.

Step 1: Outsmart google and beat the billion $ esign industry at seo Step 2: ... Step 3: IPO with Sean Parker and screw the winklevoss twins

So essentially $1100 for ~600 likes, I certainly wouldn't be happy with that return. It's an informative piece but I think your summary is a bit off regarding the campaign a success generally.

I think by simply joining Facebook groups in this sector and engaging with the group you could've achieved 600 likes in the same timescale and saved yourself the money.


We use Google mostly for adverts, and we're showing about a $1 per user acquisition rate. That's not CTR rate, that's

clicked ad->saw site->signed up.

At almost $2 just to get a "like", I would be very hard pressed to justify that.

"we averaged $1.83 per 'like'"

"what's a 'like'?"

"somebody click the like button next to us in Facebook"

"how much money is that worth to us?"

"well, it identifies interested people in our service who we can advertise to later"

"so we now know about people who might be interested in our service if we decide to spend even more money advertising to them?"


That's about the market rate for buying likes on Facebook (the company I work for used to offer this as a service).

Targeted likes or just likes in general? There are folks on Fiverr who can deliver untargeted likes at 10 cents a pop (which might be useful for the initial social proof). No idea how they do it but I tried as an experiment and they don't all look like zombies.

That seems really low for me. The only way we could have done that would be to use low value inventory like from developing countries, but usually the advertisers had a certain demo that they wanted to reach and that raised the price.

Yeah, the demo is probably not good at all. But in terms of building up a few hundreds likes in order to get social proof for later folks, it's an option at least.

Is there an industry standard on $/like?

It seems like you can't really judge how much you should spend per like because you haven't yet calculated the revenue and profits per like.

I've heard anywhere from .50 to 1.50, but you can sell them for anywhere from 3-$7 if you're an agency :-)

And now the post on HN about the ads will generate far more impressions for SignNow than the ads did.

Thats the follow up article, expect it in the days to come.

But then... THAT will be on HN. Then what?

An indefinite loop of follow-up articles, and every single one will be posted on HN. :D

Maybe a HN post on posting on HN about Facebook...HAHAHA

I took a look at their ads and still do not have a clue as to what SignNow is about. Still scratching my head as to what they were advertising..

Same here. I think the results are very bad. They say it's not about the results but about branding. Well to be honest I can't remember there logo already though I read the article..

Great experiment! Glad to see you got some engagement.

As someone who has run FB ads since the beginning though, I would disagree with one thing: in my experience over the years, people are ok with leaving Facebook, so long as you have a very nice, clean, easy to understand page with a clear value proposition.

My suggestion going forward is this: 1. Take each winning ad and make 3 more variants. 2. Take each interest you were targeting, make a separate campaign for each. 3. Place all variants of your winning ads in each campaign.

The purpose of this is to find which actual interest is responding to your product. Once you find that out, refine it again by breaking up location, age, gender, education, etc.

Get really granular over time. The end result is very good ROI and a solid way of making money on FB :)

Replying to myself here to give some info on my experiences with FB:

Here are my stats on the last week for one campaign where I make about $35-$50 on average per lead:

- Spend: $3,345.30 - Impressions: 6,305,765 - Clicks: 2,570 - CTR: 0.0408% (average, some ads are near 0.1% and they make the most money) - Average CPC: $1.30 - Leads: 237 - Average CPL: $14.11 - Revenue: $8,703.70

That's a pretty damn good ROI. Not every campaign does as well as this, but taking the time to find your audience, craft your message and optimize will yield good results.

I agree. Those numbers are pretty good even for an AdWords advertiser. And the fact that you've been able to scale it really well is even more praise-worthy.

Do you have a blog/website that I can follow?

It stories like this that make me think I should just build an app that does automated Facebook ad optimization, these guys could have done a lot better than they did. I've run ads targeting US tech audience at 0.5cents a click.

A few tips:

1) Buy ads on a CPM basis, if you're getting a good conversion rate then CPM is a lot cheaper than CPC on Facebook.

2) Don't target a bunch of keywords together, chances are some of those keywords have great CTR and some of them have terrible CTR, running one ad targeting them all means you can't filter out the poorly performing keywords.

3) Geographic focus, concentrating on the major tech areas rather than the whole US would probably give them a much more targeted audience.

4) Their budget was too high to begin with, you can experiment on a much smaller budget to get statistically meaningful results. Once you've figured out what works then you can start throwing money at it.

According to my stats [1], the implied click through rate is .04%, so you're very right this ad group would've saved money on average [.057%] (And almost half if they had stuck with the .077% converting ad).

[1] http://euri.ca/2011/04/facebook-advertising-prices-part-2/

Nice model, unfortunately it doesn't work because of a couple of factors:

If you run CPC ads Facebook will show your ads more often to users who are "click happy" (they've said so publicly), there's also an impression among a lot of advertisers that Facebook charge a premium for CPC (which they haven't been willing to confirm or deny)

Also if you're buying CPM Facebook give you a "quality" discount, so if you have a good CTR you actually end up paying a much lower price than the initial suggested bid.

Looks like they didn't really get enough traffic to do this (and idk how easy it is to do with FB and google analytics) but it would have been interesting and probably smarter to also measure each ad's effectiveness by how many conversions it led to on their site rather than just click-throughs to the facebook page. With their "most successful" ad they spent ~$500 to have a bunch of 21-year-olds click on a picture of beer that says nothing about what they do - I doubt these people will respond to a future marketing campaign any better than a random set of the population.

In any case, the analysis was interesting enough to make the front page of HN, which I'm betting will bring them a lot more than 139 visits. So well done with that.

Most would agreed that campaign did not performed, instead aim to achieve CTR > 1% and Bid < $1.

Facebook ads is used widely by affiliate marketers, you could spend some time on blogs like MrGreen ( http://stackthatmoney.com ), or affiliate marketing news aggregator ( http://affbuzz.com )

I would change the following if my objective is to get more likes & users.

- Target person, i.e vc, developers, etc instead.

- Limit and test the age groups.

- Target Tech events.

Bids will be cheaper & CTR will be more targeted. If it is not CTR, KILL it. Personally I like what you are doing at signnow, much easier to use than those I used when opening up my trading account.

Maybe you want;

- Add the Facebook icon or something to link to your Facebook page in your website.

- I am kinda of surprised that there's no YouTube explaining signnow, considering doing a hipmunk-like video explaining what signnow does.

- If you have existing customers, do a video case study what it does your customers and how signnow could come in handy in other industries or applications, etc.

re: the hipmunk video, Are you talking about a GrumoMedia video? http://www.youtube.com/watch?v=6teBPUgz4Y8

ya.. that's the one!

"Facebook ads is used widely by affiliate marketers, you could spend some time on blogs like MrGreen ( http://stackthatmoney.com ), or affiliate marketing news aggregator ( http://affbuzz.com )"

Ack...stackthatmoney just screams affiliate coaching scam. Affbuzz still has the A4D meetup in SD advertised on their site which was back in march and only has links to a few of the decent blogs. Seen way to much of that sort of stuff in my years as a full-time affiliate.

This looks ridiculously expensive for such low traffic numbers for a free app

Regarding the ad the worked best, I couldn't tell what the heck the image was, and I'm curious - the image was interesting. Looks like a spiral of some sort. (I didn't see it as a stack of beer until I looked back after reading what it was.)

I suspect the higher click-through rate was more about curiosity about the image.

I have to admit, I've personally clicked through a couple ads for precisely this reason, only to immediately hit the back button upon realizing it was just an ad.

And the next time I see the ad, I do it again.

It may be effective at getting click throughs, but the retention rate likely leaves something to be desired.

a point for improvement: facebook's auction system is quite aggressive in picking its favorites. Having that many ads in one campaign, especially one with such a low daily budget, will cause one to get nearly all the impressions and take over.

You also can't stop ads after only a thousand or so impressions. Let them run for 50k+ before cutting, as an ad that performs well will take a decent amount of clicks for the bid to fully come down.

Facebook's own guide on A/B testing your ads suggests creating a separate campaign for every single variation, in order to avoid this issue. I've still found that just 1-2 clicks causes one variation's impressions to go through the roof.


Site is down. Does anyone have a mirror?

too many images, just brought it up on the CDN to deliver the content better.

servers upgraded, sorry about any slowness.

Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact