We Wasted $50K on Google Ads So You Don't Have To (2019)
indiehackers.com
indiehackers.com
The comment I posted on the IndieHackers page:
------------------
The landing page is too complex. Like what does "Full stack adaptive delivery" even mean? I am sure 90% of your paid visitors are just bouncing because that landing page tagline is alien to them. Dumb it down. Make it simple.
Surprisingly, the description in the Indiehackers page makes so much more sense than the one you put up: "File-system-as-a-service that does uploads, storage, and media processing for Web and mobile apps, so you can ship products faster and scale them painlessly"
If you told me that the first time I would have understood your value proposition. Don't get too fancy with your taglines. People don't have time to understand what you are saying. People don't like fancy terminologies except for what is popular. There are too many jargons already. Don't complicate it further.
Instead of "Full stack adaptive delivery" just try: "File-system-as-service". Instead of "Serve ultimate UX with better images on any website. One script to rule them all." just have: "Ship products faster with better images on any website". That's it. You will get 50+% higher conversion rates with just this one change.
I've been known to complicate descriptions on my OSS packages but still achieved moderate success, so I'm thinking on how to convert this knowledge into revenue and these comments are super helpful.
Question: how do you manage to both keep it simple but relevant for SEO? Isn't SEO a lot about keyword stuffing?
Thank you. Just sharing what I learned from the masters in the field. The best of which is Isaac Rudansky. Marketing is made unnecessarily complex these days. It is not rocket science. Anyone can do it.
I have a cynical theory for that too!
The reason marketing is made complex is because it's not just marketers pushing products to the rest of the world, but also marketers pushing marketing tools to each other. Just like a marketer may exploit my fear for health of my kids, or desire for social status, they can exploit other marketer's fear of low campaign success or a desire for being a premium brand. Etc. So the industry gets increasingly opaque and full of voodoo, because the more complexity can be built in, the more people can make money without delivering any value that could be reliably attributed to them. Everyone buys each other's bullshit, money changes hands, everyone is happy, little actual value is created.
SEO should be as natural as possible. Keyword stuffing actually has a negative impact on your Quality Score (if you are running Google Ads) in the long run. Do not forget that Google is always optimizing its algorithms. What was possible before isn't possible now. And what is possible to do today isn't possible to do in the future. Don't start by gaming the system. Start by doing it as natural as you can. Optimizations can always be made later on (if required). From my experience I have seen that SEO works best only if it has relevant content tagged along with it.
Sometimes, you might be in a niche where you don't even need to do any optimizations. And most products/services fall under this category. Just like you do not prematurely optimize your software, there is no need to do the same anywhere else. Even marketing.
Natural looking to googles algos, not actually as natural as possible.
There are highly effective “unnatural” SEO strategies, for example you could just generate fake bounces on your competitors pages using proxies.
> Don't start by gaming the system. Start by doing it as natural as you can.
This is good advice. You can always game the system, but you probably shouldn’t start out by doing so.
Gaming the system can be incredibly lucrative, but also very difficult. Unless it’s the primary focus of your business, you probably shouldn’t even try.
That's not unnatural seo it's negative seo.
"Easily implement our Bi-Directional CDN© to personalize on-demand adaptive delivery of UX-relevant content. Engage your target audiences"
Slowly backs away
I think flipping it around would be even more impactful. Start with Why ;)
I agree the what often works for me, but if you want mass appeal it appears to be the Why that hooks people.
I have a file system, it comes included in every/any os that matters. How does filesystem as a service help me?
File system as a service doesn't tell me what business problems can be solved with this.
Even if I was squarely in the target audience, if I saw that, I wouldn't necessarily know it applied to me.
Maybe I'm wrong about that.
When you say your service offers a cloud-based solution for delivering orchestrated micro-services powered by AI and ML, you're not targeting technical people who know this is bullshit. You're not even after business people with a real problem to solve. You're trying to isolate out-of-their-element execs with more budget than brains, who are impressed by jargon they don't understand, and who have a KPI to "undertake synergistic partnerships with hip technical solutions providers."
1: https://www.businessinsider.com/why-nigerian-scam-emails-are...
Let's say 50% of people actually read past the tagline. People like you might read a bit further and realize they don't need it anyway. People who do need it might not read further though and you'd lose that sale.
Are these people searching through Google for a company like yours though?
It's worse than that. And it's my pet peeve about many startup landing pages these days. It's not like people don't have time to understand - there's nothing there to understand! "Full stack adaptive delivery" is a near-meaningless phrase. It can be construed to mean just about anything. It would fit just as well on a logistics company page, or on a sticker on the side of an ICBM.
I wish people would just say what they actually do.
So many times I have to spend ten minutes digging into what some of these companies actually do.
It’s highly frustrating. Give some examples of what you do. Explain it in simple language. Don’t rely on conceptual language because it doesn’t tell me what your offering does.
In my university years, I used to invent such ridiculously overblown phrases for simple things I did, as a form of mockery of corporate culture and my general pastime. But at some point I did realize that these phrases are hashing functions - like the ones you use in a hash table to put objects into buckets. So for a particular thing I do, say "adding colors to terminal applications", I could invent a bunch of nonsense phrases - "enriching the user experience of advanced software", or "delivering visual artistry to professional digital media" or whatnot. It was fun going in this direction, and we'd have a good laugh - but a person seeing just the output could never arrive at my original input, "adding colors to CLI apps". It was one of infinitely many things I could hash under the same phrase, and they could never know which one I did.
So in my eyes, if you're trying to explain something to someone, then using these phrases is essentially equivalent to taking MD5 of what you wanted to say and pasting that hash.
People try to summarize their thoughts, but if they are not good at it the summary ends up being "We're pretty cool, not like those other non-cool people, and we stand for good things". And then, having arrived at that thought, they express it using the words they have heard before in similar circumstances.
And boom - pretentious mediocracy seasoned with hollow platitudes!
"We revolutionize Enterprise Data Quality with next generation AI in the cloud" sounds better to employees, investors, and in some cases buyers
I'm going to start a business selling a quantum leap in synergy paradigms. Begin with the tagline, move on to a landing page, and after that's nicely done, get a group of people together to brainstorm what the company actually does.
Sounds like WeWork. If you've ever seen that Neumann/Kutcher interview, that's exactly how it went.
Salesman: "We provide win-win situations and customer-focused solutions."
Dilbert: "But, what is the product or service that you actually sell?"
Salesman: "We don't sell, we partner."
After that all they care about is getting you to agree that the demo looks good.
After that all they care about is getting you to agree that this product could be somehow useful in your company
Then they pounce. They ask when you are going to sign the contract because if you don't do it ASAP you will lose the discount. If you try to say that you don't want it, they start with the guilt trip. "I thought we had a deal. I have already booked my trip to come see you next week."
I feel bad shitting on a startup, but reading their docs I'm struggling to see what value they add beyond abstracting S3 and Akamai.
I guess there's a value add there if you don't have a person on board who can configure S3 and Akamai.
"Arent there other companies working on fraud on the blockchain? How are you different?"
"Our team is amazing and is made up of myself and my founder who have 3-4 years of experience and need to fill it with 8+ year experienced engineers because before then no one is really useful here."
Me: "..."
OR
"Do we really need another machine learning solution to classify documents? What is different from you and your competitors?"
"Well, we think everyone else is doing it wrong and that we've found the perfect solution."
"What is that?"
"Myself and my cofounder are far smarter than any of our competitors".
OR
"How do you talk to candidates about salary?"
"Well, we don't give any actual numbers, but we share our compensation philosophy doc, which ive linked here".
"Right, but can you afford these people? why do you expect them to take a huge pay cut to make you a millionaire?"
"The right person will have faith"
There's always an answer that feels like it was fed to them by a VC who knows it won't really work, but is just hoping and praying.
Because you are asking a meaningless question. Firstly, they are no experts on their competitors. Secondly, how can anyone even be, unless they have used all of them for a long time? Who can do that comparison??
But I also have a lesson from my former decades on earth: if lots of people are doing something that seems crazy to you, there is probably something you don't know. So could it be that there is some reason all these startups are so maddeningly obtuse? Other than just being bad at communication, I mean? Is there some incentive to not actually say what you do?
Think of it more like the well-appointed lobby of an office headquarters. I don't think anyone would expect the lobby of an office building to explain what the company does. It's just supposed to look nice and welcoming.
This makes sense to me. You generally don't see uninformative homepages from consumer software-as-a-service or direct-to-consumer companies, for instance. Airbnb and Airtable both seem to immediately explain in fairly clear terms what they do, for example.
Even massive software companies with a huge range of products that self-serve to business customers have pretty explanatory homepages. Salesforce is a decent example.
I think the home pages that get these sorts of complaints are either companies that don't have product-market fit (and thus don't even know what they are yet), or companies that have long involved sales processes with large enterprise clients where they probably tend to negotiate highly-customized services. An example of the latter that comes to mind is Splunk, whose website uses a lot of buzz words and seems to basically be saying "trust us, if you're a big company with a lot of data, we can do anything you need regarding your data."
I'm a software developer. I have no idea what they are selling, based on that. Is this some sort of souped up Google Sheets competitor? If so, what does it do better? Is this a team workflow management tool? If so, why bring up spreadsheets and databases? I suppose some teams may try to organise their workflow using spreadsheets, but this far into the 21st century, most people use tools like Jira or Github Issues or Monday.com or one of fifty other tools. Is this a software development tool? If so, errr, what?
I literally haven't the slightest idea, after reading their homepage, what problem this is trying to solve, what solution it is offering, and what the target market is. Am I a potential customer? I haven't the slightest idea. Can this help solve any problem I have? Maybe, but the homepage goes out of its way to refuse to answer this question.
Airtable has even targeted ads at me. It's not like I've only spent two minutes looking at their site. I still haven't the faintest clue what they do, let alone why they are better than competing alternatives.
I think that is why it is too generic for you. You already understand the underlying concepts of a spreadsheet, a database etc. Their target audience doesn't, but they know they need something _like_ that.
I think Airtable is in the no-code/low-code camp. Business people that have some technical ability and used to write Excel macros, can build a tool for their custom business process. Just like they did with MS Excel (or even MS Access) before.
And those people react to the buzzwords. "Oh, it can do a little bit of everything in one tool? And I can get it done without talking to the IT department? Book it."
I think you are too far ahead in your abilities to see the value it provides for people that rely on developer time to get anything technical done.
I'm not sure why you don't have an idea of what they are selling, unless you just don't believe what their website says. "Part spreadsheet, part database" really is an accurate explanation of their product. It's basically how I would start any explanation of the Airtable product.
The rest of it ("entirely flexible, teams use Airtable to organize their work, their way") is certainly market-speak, and doesn't give much more information, but it also doesn't obfuscate anything.
I personally wonder if they used the Startup Generator in earnestness. [1]
But yeah, I'm a senior techie, and often involved in potential procurement discussions, and sweet Jesus, if you want us to give you money, give us some goddamned concrete facts. You synergise enterprise cloud offerings?Oh, you mean you have a templating language that tries to generify Terraform and CloudFormation and does both badly.
I sometimes feel like people in ticket clipping businesses like this (their CDN offering is, um, Akamai, but you can make pictures grayscale using their DSL because that's easier than using a photo editor?) are scared of saying what they actually do because then you'll realise that they don't do much for the money they're asking.
Maybe pay that $50k to a consultant to fix your copy.
Granted there was some mob subtext to that line but the world would be a better place if people strove for that level of simplicity in their messaging.
As someone else said in this thread, do you optimize for venture capitalists instead of customers?
$50K of lessons learned, summarized for you Let’s recap our lessons learned:
Narrow down your extended keyword set and focus on the keyword groups you have polished content and landing pages for, especially when you have a complex product.
Use an awareness ladder to inform keyword segmentation by purchase stage but revisit it often to validate and adjust.
Do not use the same landing pages for different steps of your awareness ladder.
Do not wait for leads to become paid users to decide on the quality of paid ads campaigns: focus on quick metrics and tailor experiments to one step of your funnel at a time.
Take a test-and-learn approach with small budgets to quickly fine-tune campaigns, focusing on page quality and clickthrough rates.
Run tests to optimize page content, which will reduce your cost per click.
Once you find your winner, you’re ready to go all-in. Now’s the time to pass it off to an agency to scale things up, if you are contracting out campaign management.
Also, should landing pages differentiate based on the query string ("?v=s" here), or be hosted on a new path?
I will break them out into new paths, thanks.
Landing pages should preferably be hosted on its own path. Search engines index paths not the query string.
Also, think this way: if you have multiple features, you can have a landing page per feature explaining that one feature in depth. Then you can have your Ads target that one feature page that everyone wants. This helps when you introduce new features tomorrow. Instead of cramming everything into a single landing page (your homepage).
I'm not sure I've ever searched for something and got a different landing page as a result (though maybe I just didn't notice), so I didn't realize this was something people do.
> I'm not sure I've ever searched for something and got a different landing page as a result
That is probably because Google was shortening the destination URL. If you are running Google Ads you can set your own URL destination even though the actual landing page might be something totally different. You can check it out if you like by searching for something and focusing on the Ad. You'll typically have the landing page different from the Ad URL.
BUY THE SITE
$999,999.98 Per Year
1 GB
What do you care? You own the site.
Still no support
BUY NOWAnd yes, the tagline I mentioned may not be the best but it will definitely get him more conversions compared to the one he currently has. At the end of the day conversions matter the most. No sales, no business. Doesn't matter how much analysis you do.
My first thought was that someone is going to upload a 43 meg animated GIF and then complain that it's loading slow. Not a comment on your product, just past experience tells me it will happen.
If those people are not your target market, don't give a shit about what they say related to the landing page content. You're not selling to them. If they don't understand the terminology used and they are not potential clients, you don't have to make a "less complex" landing page or change the wording just to make them happy.
The idea behind a tagline is to capture the user's attention. You can build up on the complexity as the user goes further down the page. But to start with a complicated tagline will cause your users to quit.
If someone is selling mangoes in a market and he shouts "Mangifera Indica for 5$" will people purchase from him? He is still selling Mangoes and dare I say he is more accurate (because the scientific name for Mango is Mangifera Indica). His potential customers will be those who are botanists because only they understand the scientific nomenclature. But those won't be his real customers too. What if the botanist hears the scientific name but his/her brain ignores it as the brain triggers only in certain environments (like in a lab)? What if the botanist is not interested in Mangoes even after understanding it? As you can see, your potential customers are narrowing down with every filter applied. This is exactly antithesis to "selling to as many as possible".
But if he shouts "Mangos for 5$" he attracts everyone interested in a Mango, including the botanist.
And I don't want to sound nasty but no technology is so advanced that it cannot be simplified. Also don't forget that people can smell bullshit from a mile away. That is what I have learnt from experience. We Engineers like to overcomplicate and overpromise stuff that will only come back to bite us later. Just follow the KISS principle: Keep It Simple, Stupid.
The "Adaptive Delivery" is the new technology we're currently testing copy for. Actually, we can also discuss the best explainers for "we analyze user context and tailor media content accordingly with our Image Transformations CDN API, serve it from Akamai." The full-stack thing is there to "show" we're leveraging the complete Uploadcare pipeline for just one line of code implementing the "adaptive behavior".
You provide your image URL It gets fetched to Uploadcare via reverse proxy Once it's there, it gets to our storage and is cached on CDN layers Then we analyze the page layout and tell the API which image version we want exactly API produces the version It gets served personalized to your every end-client session
I'm guilty of this as an engineer in a startup. Engineers thrive on the details. But users don't care too much. They just want the service.
> Lesson #1: Test and trim keywords sets before hiring an agency to scale things
This isn't how agencies work. You did the keyword research and testing yourself and then paid some one else to do it again.
I work in an agency, I won't work with you if you do this, because you will get the same results and blame me.
> Lesson #2: Focus on page quality and CTR when doing paid tests
Nope don't do this. When testing ads you should focus on the ads you are testing.
Make as many as you can, test, review and reduce to the winning ads then repeat. When you have found the ads with good CTR it's time to start working on page quality (making your page relevant to the ad).
> Lesson #3: Stick to keywords that you have landing pages and content for
Again no don't do this, write ads to test your keywords build pages for the ads that work, not the other way around this isn't SEO.
> Lesson #4: Don’t assume organic conversion rate will hold true for paid
Different ads target customers at different stages of the buying journey, ads don't dump to the top of your funnel they dump to landing pages designed to convert that demographic.
> Lesson #5: Analytics will save (some of) your bacon > Lesson #6: Revisit your awareness ladder often to validate and update it
Winner winner chicken dinner, some good advice.
Regarding this:
> ads don't dump to the top of your funnel they dump to landing pages designed to convert that demographic.
What do you mean exactly? Isn't a landing page the top of the funnel by definition?
This user-centric idea of a funnel is different than the website-centric view of the funnel where you start with a landing page and end with a lead form or purchase. The phrase "Customer Journey" usually gets bandied about to refer to the former, although it also often gets bandied about without referring to anything at all.
This isn't how agencies work. You did the keyword research and testing yourself and then paid some one else to do it again.
Unless I misunderstand you, you misunderstood them. They are saying that they didn't test and trim first and paid a lot of money to the agency that they didn't need to. They should have started with a smaller set.
> Lesson #4: Don’t assume organic conversion rate will hold true for paid
Different ads target customers at different stages of the buying journey, ads don't dump to the top of your funnel they dump to landing pages designed to convert that demographic.
So do organic searches. The initial Google hypothesis was that if someone was searching they would be interested in buying, and showing them an ad right then would be optimal because they are interested in buying (or else they wouldn't click it).
The idea that conversion is better with organic than ad does indeed seem to falsify that hypothesis. But maybe not. There could be click-fraud involved, too, which can dramatically drop the conversion rate.
The agency will add 20% on the spend because they can get better results than the client doing keyword research, writing ads and testing them.
The client did the hard work for the agency basically made the campaign themselves and then paid the agency 20% to set it live.
They should have done it the other way around get the agency to setup the campaign and then "scale" it themselves and they would have been able to an extra 20% ad spend for nothing.
The second point you are right again. I mean more that the funnel should start at the point of contact with the end user (the ad) and the landing page should be built for that ad, they seem to have tried to do it the other way around.
They obviously know what they are doing as far as tactics, it looks like they learned it from SEO, but I think they have tried to apply an SEO strategy of content first to their ad campaign and it didn't work.
You can't change advertising channels so you need to change your content to fit the channel not the other way around.
It may as well have been written by an ancient alien civilization for all I understood.
Probably we’ll hear something very like this when SETI receives a signal from another star system.
Now, where are the statistical metrics and causal inference assessments?
I’m glad they clarified what “net negative churn” means, that really cleared things up.
Thanks to developers’ trust in our core infrastructure and their recommendations, we are looking for ways to improve the quality of our products by providing a service to our clients with the highest level of service for the remainder of the year.
> Thanks to developers’ trust in our core infrastructure and their recommendations, we're able to deliver the highest quality services available. We're focused on delivering your most productive platform, whatever you're doing with it. From your smartphone and tablet to servers in data centers and data centers in offices. Our solutions help you store, process and manage everything from plain text and binary data to HD video and audio files. About Us: Cisco is a leading technology company known for innovation, innovation, innovation. We use technology to create a better world. With global operations in more than 100 countries, Cisco provides the essential technologies that enable companies to compete and win in today's global economy. For more information, visit www.cisco.com.
> Thanks to developers’ trust in our core infrastructure and their recommendations, our deliverables are performing better than ever. We are further expanding our market reach and capabilities, making us more competitive and reducing cost through better processing, architecture, security and data centers. That's the good news. The bad news is, we're reaching our head-end capacity and need more help. The Rails team members that work on applications and features on a daily basis don't want to burden the engineering team with refactoring and maintenance. They know that when the infrastructure is gone, they have a new responsibility to help build applications for new players. And there is a new player, but we are still in its infancy.
> Thanks to developers’ trust in our core infrastructure and their recommendations, we have secured the Ecosystem. This Ecosystem is already in the technology and revenue channels process. We have a team working on it, including the core guys and big but skilled giants who have also invested their expertise and assets to become the gateway into our ecosystem. As more people start to use this new feature, they will ask for news. In terms of devs, we already have more than 40 developers on board. We have identified a cluster in Africa which is running our client, and have a major team dedicated to project execution and content creation. As development work is ongoing and we don't have exact numbers, we can't officially announce how many people we have on board, however we are not aware of any
They aren’t necessarily getting more new customers than old customers cancelling, but the combination of revenue from new customers and added revenue from upgrades from existing customers is more than the revenue lost from customers who cancel.
Which, btw, is exactly what they say, using standard industry terms, and it’s not ambiguous. It just wasn’t written for you, that’s all.
They are getting more increased revenue from existing customers than they lose from cancelling customers, which is very different
Churn -> people canceling their subscriptions.
Churn Rate -> rate at which they're canceling (ex: 5% a month)
Net Negative Churn Rate -> if the overall dollar amount of people moving from $50/mo plans to $100/mo plans is higher than the dollar amount of people canceling subscriptions.
So while the Churn Rate doesn't change (still 5% a month) the overall monthly revenue is still increasing. This is really hard to do and it's a good indicator that they're making something that people really want and are willing to pay to use.
New customers spending new money (net new)
Current customers increasing spend (expansion)
Current customers decreasing spend (contraction)
Current customers leaving (traditional churn)
Net negative churn means when you add up the bottom three, revenue has increased.
Another way to say it is you don’t have to add new customers to keep growing revenue.
let use_increase = (Increased income from existing customers using the service more)
assert churn > use_increase;
net_churn = churn - use_increase;
assert is_negative(net_churn);
> The best minds in any profession are never guilty of jargon, expect when they are very tired. Pedestrian minds are drawn towards it automatically and to the most frightening extent. Jargon, one could suggest, is the natural weapon of highly paid people with very little of any value to say. It is a sad and ironical comment on our society that many people feel released from the pressure to use jargon only when they have reached the top of their profession, by which time it may be too late to change one's habits. , however one might wish to. Ambitious people, still busy climbing the ladder, may well consider it professional dangerous to use straightforward language. One therefore has the paradox that only the person who has finally arrived, with [their] reputation secure, can afford to be simple and jargon-free. Lesser mortals appear to need their jargon, as a membership-badge of their profession. They do not have the confidence to face the world without it.
Do I escape the tendency towards jargon? No, not by a long shot, but pieces like this are a good reminder as to why it is worth the effort to keep trying.
And so just like you wouldn't remove the jargon from a medical journal not sure why you would want to in this case.
I'd say pretty much everybody's :)
I saw one where they where talking about keyword density and counting the names of the <div> elements.
The name of the consultancy is redacted
Might be worth watching some of the YC Startup School videos to get up to speed.
I've no doubt that if you're unfamiliar with marketing that these terms are new (in the same way that if you're a marketer and reading about database indexing you're going to have to do a little work to figure out what everything means.)
This is important to me not because I've any stake in this (I've no affiliation with Uploadcare) but because I worry that the knee jerk rejection of any kind of marketing by developers really hurts startups and bootstrappers.
There is definitely blather in the marketing world [1], I'm just saying this isn't it.
1 - https://www.cbsnews.com/news/pepsis-nonsensical-logo-redesig...
They were saying they should have held off on targeting those more specific terms until they had those matching and more specific landing pages in place.
You can’t say “hey ordinary person I’ll save you wasting $50k on AdWords” then instantly drop straight deep into the most arcane of mumbled, cryptic, faded rune digital marketing magic speak.
Refer to the comment below from HN user 83457 who explains same but in English.
The writer establishes they have some baseline analytics and understanding of demand generation prior to embarking on paid.
Then they embark on paid.
Then they tell you what they learned.
FWIW for any organization new to paid, this article will probably save you some 10s of thousands in spend if you internalize it properly.
I feel a bit the same way reading this article. There's a lot of talk about conversion rates and funnels and semantic cores and awareness ladders but in the end does do they have a product that people actually want?
Having bad SEO and a confusing website is bad because it could make it harder for people to buy your product but it's not going to magically make people buy something they don't want.
Also I wonder if focusing entirely on "conversion rate" is fair. There are other benefits to advertisement, such as increasing brand awareness for instance. After all look at the vast majority of real world ads, there's basically a 0% conversion rate. People don't see a Coca Cola ad on TV and immediately go the supermacket to buy a bottle. I don't see a football player wearing Nikes and immediately go online to shop for shoes.
Maybe somebody will click on your ad, browse your website for a bit, think it's interesting and months later they'll convince their boss to use it for their next big product.
Their example query of "file upload php" is someone who's looking for a solution right this second, so it's not unreasonable that they would sign up and try out the service right this second.
The reason people focus on conversion rate is because you can easily calculate spend and predict RoI. Coke can spend on a TV ad because it's a drop in the bucket but every dollar is valuable to a bootstrapped company. Exactly why you want to know what you'll get before one blow 50k on adwords.
Then when I go to the website: "Full stack adaptive delivery"
I have no idea why their adverts aren't converting, I mean who isn't into full stack adaptive delivery? So. Damn. Compelling...
this is because that's how google gets paid.
you can create a great ad that's not terribly relevant to a specific audience, get good ctr but poor conversion.
some tips:
make sure to watch your search query reports and continuously refine your [hopefully shared] carefully-applied negative keywords lists.
avoid broad match like the plague. we use mostly phrase and modified broad with good negative phrase lists that have been continously honed over many months.
don't hyper-segment keyword or ad variations, it'll become a nightmare to manage with little-to-no benefit.
also, whether paid ads are worth it is highly dependent on your market, competitors, and your typical conversion value.
because our typical orders exceed $500, we see double-digit factor returns on ad spend even though we're an established, well-known mfg name in our market.
and yes, be ready to waste some money when dialing things in - $50k in a month of pure waste is quite a lot but that same amount "wasted" on tweaking over the course of 6mo-1yr is not out of the ballpark, assuming it yields progressively increasing ROI.
YMMV
if i was google, i would make sure that the ctr was as high as possible while the conversion rate was just enough to be worth it.
there are a lot of fine-grained controls missing from the adwords backend which waste money. for example, you cannot segment iphones from android, or forcibly exclude ie11 because you have chosen not to support it any more and the site falls apart on it. you cannot prevent ads from showing to visitors who've already come and bounced. they keep expanding how much "modified broad" sucks in by making the matching looser - with little notice. and that's just the tip of the iceberg.
lots of things that can easily save you money do not exist in their backend. imo these are not oversights but are left out intentionally.
> "...a customer acquisition cost way above one-third of customer lifetime value"
Is this really so terrible? As I understand it a lot of mobile games pay more for acquisition than they'll make from the user in e.g. IAP's to scale up their player base and climb the charts. What is a normal CPA vs LTV in the SaaS space?
Also, if you pay less to acquire these customers than you will make from them over time, you haven't quite wasted 50k have you?
This sounds self-evident but we all know sales people who sell a fiction and then engineering has to scramble to (try to) turn it into reality.
People click an ad to learn more, if they are presented with a specific use-case but land on a generic marketing page they will click the back button.
That combined with the increasing crackdown by the consumer on blocking data tracking raises serious questions on the long term viability of the business models of some big players in the industry.
> A clear funnel with established conversion rates, check
Is it checked? > 20% of people hitting our landing page block Segment which then blocks a portion of our analytics. It's pretty hard to get the full picture on your users, especially those coming from ads.
For years myself and other engineers have been warning the product owners that third party user tracking's life span is very much finite - Firefox was going to block 3rd party cookies by default several years back, but walked it back because Google weren't ready for that and threatened to withdraw the money Mozilla makes from them, but it was the sign of the end times.
Now that Google can identify you with reasonable probability without using cookies, they won't hold Firefox back, they don't need to, and it suits them to roll out the third party blocking in Chromium, alongside browser cache partitioning to avoid the classic circumventions - ETags, steganographed PNGs etc. as first implemented by Safari. UA strings, often used in fingerprinting, are also not long for this world.
And of course, now that Google has a solution that doesn't require cookies or fingerprinting, Chrome is all about that privacy lol - the fact that it hurts Google's competitors is merely happy circumstance.
The amount of panic in adtech companies now is hilarious because they chose to mimic an ostrich until the browsers left them no choice - but they're still flailing around and demanding devs try to find alternatives to 3rd party cookies.
The cookie based user tracking paradigm is dead, but there's a shit ton of people, especially in DSPs and DMPs whose salary relies on them not believing that.
Luckily, the partitioning of browser caches on a per domain basis as pioneered by Safari has now made its way to Firefox (I think it's been released) and I presume will eventually be turned on by default in Chromium: https://www.chromestatus.com/feature/5730772021411840
Explained in more detail here: https://news.ycombinator.com/item?id=21469677
Prepare to be disappointed by paid traffic conversion going in. Paid traffic isn't there to convert, like you hope since it costs you money. It is primarily there for brand awareness which only works if your branding is visible.
This sounds very familiar... in fact it sounds a lot like "Respond to change instead of following a rigid plan" from the agile manifesto.
Wouldn't surprise me if bounce rates are high. Took quite a bit of reading to figure out what industries and use cases are
When I ran a brick and mortar business where we took bookings, Google Ads were imperative to us getting new business. So were Facebook Ads. When they weren't running for some reason, or weren't optimized, we could feel it in the pocket book. AND, it wasn't too expensive. I was able to pay around $300 per month and get a lot of business.
Anyway, every business is different, and I think local business owners need to consider whether Google Ads are good for them based on their own unique situation.
You seem to have learned your lessons like you said and I actually picked up some interesting things from here. I'll be saving those Google docs!
I think the biggest lesson by far to be learned from this is that working with SEM agencies can be a complete clusterf*ck. Most of the time I'd recommend working with a freelancer who has a strong and verifiable background (most can be found in online communities).
Like you said the strategy these guys usually use is to throw stuff at the wall and see what sticks. In some, if not most cases they'll have scripts or software running to automate their campaigns without a human going into the account very frequently to weed out bad search terms.
While I 100% agree with most of your conclusions and the conclusions of the people on here commenting on your landing page wording, I do have one question: did you have a good, hard look at the search terms before deciding a keyword had to be removed?
O and yes, quality score is important. However I've seen and ran campaigns with a <5/10 quality score but with incredible CPC's. Don't get too hung up on QS - if it works it works, don't question it.
https://www.getleadup.com/post/the-startup-founders-guide-to...
"o Narrow down your extended keyword set and focus on the keyword groups you have polished content and landing pages for, especially when you have a complex product.
o Use an awareness ladder to inform keyword segmentation by purchase stage but revisit it often to validate and adjust.
o Do not use the same landing pages for different steps of your awareness ladder.
o Do not wait for leads to become paid users to decide on the quality of paid ads campaigns: focus on quick metrics and tailor experiments to one step of your funnel at a time.
o Take a test-and-learn approach with small budgets to quickly fine-tune campaigns, focusing on page quality and clickthrough rates.
o Run tests to optimize page content, which will reduce your cost per click.
o Once you find your winner, you’re ready to go all-in. Now’s the time to pass it off to an agency to scale things up, if you are contracting out campaign management."
While I'm being cynical, Indiehackers could stand to not load their css in js, at least for the global stylesheet. Surprisingly the content actually renders, but the header svg and so on breaks quite badly.
A good reminder that you can apply The Lean Startup methodology (Build > Measure > Learn) to projects of all sizes from forming a new company to running an ad campaign. Ship early, test often, rinse and repeat.
Lesson #2: Don't try to boost sales just via marketing. Great product sells itself through word of mouth.
A perfect case of finding opportunity amid crisis.