Hacker News Reader

Comment by jermaustin1 | Hacker News Reader

Author here. I'm happy to answer any questions or comments. A little about this post:

I was recently commenting on an excellent Show HN for a product called Duet and it was the most karma I have ever received on HN (17 votes in 4 hours), and another respondent said I should write it up as a blog post. So here it is.

View on HN

Actually, I have another project (in progress), which is a collection of curated blogposts. The name is http://www.bestprogrammingbooks.com

It's mostly books I like, and books I've read, but it's current revenue comes from Amazon affiliates. I know affiliates are often frowned upon, but I do my best to not recommend anything crap ;)

In the payment processing section, you mention billing the customer back for the processing fees. I was under the impression that credit card companies do not allow you to pass surcharges onto the customer. Is that not the case when using Stripe, or is it considered a "convenience fee?"

He's billing the SaaS customer more, not the customer of th SaaS customer.

You aren't passing the fee on to the END customer, you are just acting as the payment processor for YOUR customer. You can charge them what ever you want. Apple, Amazon, Google all charge you 30% to use THEIR credit card processors.

So the end customer buying a $20 product still only pays $20, but you charge the store/company they bought from $1. The company ends up with earning only $19, and you end up earning something like $0.50 after you pay the portion of the fee you were charged by stripe or paypal or whoever.

I think this is likely a really bad idea for 'side projects' ... by acting as the middle man here you are taking on a huge amount more liability than you would be if you are using the (non-managed) Stripe connect model where the your customer has their own processing account.

Sure, there is money to be made here, but it isn't free and you need to manage the extra risk and liability that comes with it. What if someone discovers your platform hasn't got the same anti-fraud/verification processes that the larger players can afford by virtue of being a larger player, and decides to take advantage of it?

This, this, this, this. Commented about this issue above but then found this comment here later. I've read about side-project startups that have gone under because they tried to do their own payment stack and lost at the fraud detection game.

See https://news.ycombinator.com/item?id=10234287

--------

Since pro accounts initially cost £6 for month, it turns out that this is low enough that it won't send red flags to stolen cards. ... If there's a dispute on Stripe, there's transaction fees for reversing charges. £15.34 in fact. Since I know I'll lose the dispute, it's cost me, £21.54 to allow some shithead to use JS Bin as a stolen card testing facility.

--------

Another good story about reducing test fraud: https://www.candyjapan.com/how-i-got-credit-card-fraud-somew...

> "..Sure, there is money to be made here, but it isn't free and you need to manage the extra risk and liability that comes with it."

What's behind the EXRTRA risk? Couldn't he just simply pass over the liability to Stripe who's the actually payment processor?

You can, and we did.

We had an accounting software that was used in some pretty big Martial Arts schools, we handled their payments through Authorize.Net (stripe didn't exist yet), but in the contract, we told them a reversal will cost (making the numbers up because it has been so long) $25 + the transaction and a bounced eCheck was $30 + transaction. These fees were directly pass through, we didn't try to profit on them.

That said, the number of fraudulent card/check or bounces was 0.

We were the payment processor for a business, not working with skeezy customers directly (like Duet is).

If there is fraud, you bill that back to YOUR customer. Just like Stripe bills you for a bounced check or a reversal, you bill your SaaS customer.

The fraud I'm worried about isn't from a bona-fida customer, it's from a fraudulent customer.

BadGuy uses your service to sign up for BadGuyDojo, and attaches a stolen bank account. 10 "customers" sign up with stolen credit cards, process payment, and the money is gone along with them. A month or two later the credit card companies come looking to you for the money.

The experience you are talking about wouldn't really jive with the real world scenario which this would be deployed in.

I'm a small business owner and you are my client. You and I have a relationship. You aren't a random BadGuy. The likelihood of you purposefully bouncing a check or committing fraud is next to nil. In this scenario, similar to what Duet is doing (facilitating communications and payments between known vendor-client relationships) the SaaS that processes the payments is pretty safe from fraud.

Now, I would never roll this out on any of our SEO services. When we had credit card payments on the earliest version of Linklicious, we fought over $1200/mo in charge backs from fraud or "fraud." After nearly 3 months of our mailbox filled with chargebacks and complaints, we switched to PayPal and only once did we have an obscene amount of fraud (hundreds of signups from stolen French PayPal accounts) at which point we blocked the country of origin for a week and the problem disappeared.

You have to be intelligent no mater what payment method you chose. There are negatives on both fronts.

Very interesting article. We are also working on a small side project, called https://www.MetaGrader.com

We are aware of the aesthetics of the site, and working on that, but we have some discussions on how to monetise it; should we go for a credit system (with weekly reports about your site, historic information, automated improvement tips), should we go for selling expert advice, or should we focus on other things, like advertising, reselling services like pingdom/gtmetrix, etc?

We would value your opinion in this!

This would be a lead in to a back end product.

Your back end product could be something like "managed customized cloud hosting" (a.k.a. 3 $5 digitocean droplets) $85/mo. I did something like this with a wordpress hosting product we launched. $50+/mo. We just used a specialized hardened image on digitalocean. The lowest product was a single $5 droplet for the wordpress install and a $10 droplet for the MySQL box.

Or hold back some of your recommendations, generate a report with contact information so they can get in touch with you and set up a service agreement giving them EXACT things to change. Try and set up a retainer for 5hrs/month @ $200/mo then tack on specialized hosting!

Much appreciated for your feedback. I think we'll go for option 2, especially for the recommendations. Also, one of our services will be continuous monitoring of your ranking and performance, so that any changes done to your site are reflected in a timely fashion. This could match with a subscription service.

At the moment, I think we don't want to recommend specialised hosting when we cannot prove it will benefit their site, but we might add a host comparison to get better data on this.

Thanks again for the help, your feedback is much appreciated!

I wish I could mod you up.

As I do lots of hobby projects, I like your advice and will definitely try some of the pricing approaches.

Keep up the good work!

Thanks! If you have any questions you can post them here or if you want to talk directly, you can twitter at me (is that the saying?) @jeremyboyd

Hey, thanks for writing this.

What do you think of MNMN.io? That's one of my side projects.

From what I can gather on you page, you are offering a service that reads an article for a person then gives them a brief summary.

This landing page is too busy with the animated backgrounds as you scroll.

From my short look, I think you will have a hard time finding that person who has no time to read an article and instead wants to outsource that to you. You would do better if instead of having a human summarize the articles, utilize one of the many NLP Summarizers out there (they aren't great but they scale). More about that here: https://en.wikipedia.org/wiki/Automatic_summarization

You can charge users for credits, then base how much an article costs on its length.

For a product like this it is always best to hide the true cost behind credits, then charge tiered prices for the credits.

Buy 100 credits for $7 or 1000 for $25, etc.

You will then want to explain WHY the person NEEDS this, and probably in autoplaying animated video with a professional voice over actor reading a script that is also close captioned in the video (test the autoplay though. some audiences convert like crazy on it, others bounce).

Then have 3 product offerings, NLP, Human Written and a Voice Mail type MP3 recording of the summary.

Most people will go for the NLP, but some will pay for written, and you can out source that to various crowd sourcing platforms for about $1. For the voice mail you will probably have to pay $1 for the summary then $3 for a person to read it.

Credits should be purchased on a subscription plan for a discount over the one-time purchases, and you should offer a few sample outputs and get testimonials from people.

That should get you started! If you have any other questions, let me know.

Thanks for the suggestions. I want to try a human-powered summary first, and see results.

I agree that credits might work better, but I love transparency and therefore I'll also stick with money :)

For sure, and always do what works both for you and your customers. I know that credits have worked for me in the past when the pricing doesn't make much since in US$.

>For a product like this it is always best to hide the true cost behind credits, then charge tiered prices for the credits.

With due respect to jermaustin1, please, people, do not take this bit of advice. Using credits or other pseudo-currency is a standard dark pattern used by casinos and free-to-play games to extract more revenue from customers by obscuring the amount of cash that a user is paying for a given action. See, for example, http://evilbydesign.info/greed/money-to-tokens/.

It signals that a business is willing to take advantage of human cognitive blind-spots to get additional revenue rather than being purely focused on delivering good value for money.

For people who disagree with RandomOpinion, care to share why? I'm struggling to think of any valid reason for using credits.

It does allow you more flexibility with pricing. As an example, iStock sell photos that cost 1 credit or 3 credits depending on how fancy they are. But the credits cost a variable amount depending on whether you bought a few credits or hundreds. So they can offer a tiered a la carte pricing structure that also has bulk discounts. Hard to tick all those boxes without credits.

Multiple companies I've used simply offered me a bonus in $/€ when I made a larger deposit, say, paying 20€ gets you 20€ in your account, but 40€ gets you 44€ (bonus of 10%).

It's currency independent. Instead of having to provide a different price for each country/currency, you can just say it's a certain number of credits and simplify the look of your store for it. That's one reason why it's common in virtual stores in games (it is still a dark pattern and can be used to take advantage of people, I agree, but it also has its benefits).

Everyone around the world can discuss an item and be like 'it costs 50 tokens for blah' and do comparisons between items in the store based on simple token amounts, but actually be paying different costs (and see different amounts) for it with their own currency.

To use programmer terms, it adds a layer of abstraction on top of the service, with all the associated pros and cons that goes along with that.

I'd like to see some examples of the summaries. If you've performed a few already, it's literally no additional work to post them.

You might want the units you charge by the word summarized. A possible issue might be highly technical texts that require specialized knowledge. You could charge more for those.

Your business model is basically what I call 1:1 - One person pays for a summary and you perform one summary.

It might be more profitable to do a N:1 - A bunch of people pay for the same summary and you perform one summary.

This is how SparkNotes and CliffNotes work. Imagine selling "The top ten daily Hacker News posts/comment threads summarized daily."

Even better would be to offer summaries of things many people want that are relevant for longer periods of time (so you can sell to more people). A site that summarizes famous speeches would be one such example.

Sure, you can see some summary examples here: https://github.com/simonebrunozzi/MNMN

Those are great suggestions, especially the speeches.

To add some bikeshedding... One more opinion: get rid of that swooping stuff from the sides. You are trying to sell "easy to read" -- don't make the summary of your services annoying to read.

Edit: also, you mention transparency to prefer money to credits, but I don't see prices.

Good point... Thanks!

Thanks so much for taking the time to write this up!

Could you tell me what tools or services you use to automate the emails, and detect which ones to send based on if users are using the service?

Thanks

We used InfusionSoft and a inhouse email marketing tool when that one became too cumbersome. This can all be done now with ConvertKit and Intercom.io, but those tools were either in their infancy or not available at the time.

Bookmarked this, for later reference - I've got a few small webshops that are under performing - the email follow up plan seems like a really good tactic. Cheers!

Déjà vu! I wish more people would communicate via blog posts ...

Reply on news.ycombinator.com