Hacker News Reader

Comment by siftrics | Hacker News Reader

>Reduce decision fatigue for customers.

I am a huge evangelist of this piece of advice. I'm the founder of a startup that offers ONE price --- and free trial credits, of course --- for each product. I think it has helped us tremendously.

The amount of startups I see that use the 3-column pricing plans (or worse!) is mind-boggling. Perhaps there is empirical evidence supporting complicated pricing. But I've never seen it. I specifically went with one price per product because I think complicated pricing plans negatively affect sales and sign-up rates, not to mention operational complexity!

If you're interested, here's our pricing page: https://siftrics.com/pricing.html

View on HN

3 column is used for anchoring you to get to the desired pricing (the middle one).

https://www.nickkolenda.com/conversion-optimization-psycholo...

Also, if you can change the framing so the decision becomes 'which option should I go for?' instead of 'do I want to buy this at all?' then that's a valuable bit of sleight of hand. But it comes with the risk of decision fatigue

While this might be true from a sales perspective, I'm actually annoyed when there is a product I want to buy and then I see these tiers.

But how many times have you said "I was going to give you money, but now that you're giving me options on how much money to give you, I'm not going to give you any."

Wouldn't it be more like "I was considering giving you money, but if you're going to make me work for it then I'll reconsider"?

This is a valid point. I've gone to sign up for a SaaS before, read through the features list and thought it was great, and then realised they've split all the features up among plans and I'll need to pay more than I was expecting to. Obviously this just means that the highest tier plan is what they think all the features together are worth, but I often find that for smaller SaaSes, the desire to have 3 plans shapes how they do their pricing, so the highest "all inclusive" plan is more expensive than it otherwise would be if there was a single plan.

It's more like yjftsjthsd-h said in their comment (https://news.ycombinator.com/item?id=24548071): They suddenly make me work for it.

Sure, this might be the worth of the features, but then I might get frustrated: "I want features X and Z, but they are only in the most expensive tier. In the cheaper one I can get just X, or X and Y. So now I have to pay for X, Y, Z although I only need X and Z."

Another example: johnreed fitness gym offers three tiers. They are basically the same, but give you more flexibility in where you train. If you are a member for one year, you move up one tier for the same price. It's not clear whether or not my contract would get cheaper if I start at the highest and after a year move to the price of one lower.

So basically they made me work and think a lot just about pricing for some service I actually want because I don't want to think about it. Which made me choose another gym altogether.

3 columns. Left-most = maximizing TAM. Center = optimizing revenue. Right = maximizing revenue on strong fits.

TAM?

Probably total addressable market. https://en.wikipedia.org/wiki/Total_addressable_market

Thanks for sharing your company, https://siftrics.com

I'd never heard of your company but this service looks GREAT.

I wish you financial success.

Question: your price points are sooooo low (which is great for the consumer) but I get concerned it's too low for your company to stay viable. Is that a concern? Please don't take this an me knocking your company. I really love what I've seen on your site.

What exactly is "soooooo low" about $0.50 for a 1000 pages?

If your customer is someone who previously had a team doing manual data entry, this pricing certainly leaves a lot on the table. I would imagine most customers being happy to tolerate a 10x increase in price without breaking a sweat, assuming there's no one else in the space offering a cheaper solution.

Said 10x increase would put them at ~ the same price as the amzn and goog services the homepage mentions, so...

I can't imagine a use case where I'd be willing to go the hassle of paying $0.50 for 1000 pages but wouldn't pay $2.50 or more for those same 1000 pages.

It would take me probably an hour or more to scan 1000 pages. Given that, I don't see a significant difference between spending "an hour plus two quarters" and "an hour plus ten quarters".

First of all, thank you!

The point of the pure text recognition service is to draw business; not to make a profit. Without the "Advanced Features", the service operates at a slight loss. However, the profit margins and total volume we do on "Advanced Features" covers the loss and turns a profit.

On the Hydra (Documents-to-Database) side of things, I think we charge enough. Margins are very high.

Much of the time I know I need to do 10,000 API calls and do them in the next week. I don't want to get "X API calls a month". I love the simplicity of AWS pricing where you pay for what you use -- except for the "free use" part which leads people to spend $5000 worth of time to save $5. (I would be richer than Jeff Bezos if I had a penny for every time somebody did something stupid or wasteful to stay within the free use tier.)

> if I had a penny for every time somebody did something stupid or wasteful to stay within the free use tier.

You make a good point. I suppose it doesn't apply in quite the same way here, since we offer lifetime free credits --- they don't refresh each month. 1,000 free pages, then you have to pay.

Lifetime? Really?

I'll still be able to use this service in 2075?

Do you expect anyone to believe that?

Lifetime of the company, presumably, not lifetime of the user necessarily.

Thus it's meaningless. Put "lifetime" or "unlimited" into a pricing plan and the one thing I know is that neither I nor the vendor can think rationally and honestly about it at all.

Thus it is a reason to look for another vendor who shows signs of realistic planning. If I can pay $50 for something that costs them $25 and it is obvious that it is roughly like that, I know I am partnering with a 'sustainable' business.

Seems like pretty realistic and sustainable planning to me, they give you at most 50 cents of credit to see if it works for you without a forced time limit.

That's why it's just better to have time-based pricing, like a monthly subscription. As long as you keep paying, you'll be able to access it.

You might have a point with general lifetime pricing plans, but in this case, I think the usage above of "lifetime free credits" was perfectly clear to most (and is in fact, limited).

If you look into the fine print, "lifetime" or "unlimited" really means "until we change our minds".

I assume the free credits are for trialing and testing.

Lifetime is never the lifetime of a person. It's the lifetime of the account (limited by the lifetime of the company).

If a company gives me "free X for life", I can't expect X if they declare bankruptcy. If they get acquired, the acquiring company will probably be within their rights to cancel my free X, and this has been done time and again. You can try to die on the hill of "lifetime means my lifetime" but I don't know what you expect to gain.

This is the first time I've seen someone call AWS pricing simple :)

It is compared to the "pick one of three plans none of which make any sense".

I know investors like recurring revenue, I sure do.

For many things I have no idea how much I am going to use them in six months if at all, but it seems they feel compelled to price something at "$X a month" even if I just could spend $Y on it this week, solve my problem, and maybe buy some more later. I am thinking in terms of $Y.

There are just so many monthly bills out there that the only service anyone needs now is one that cancels all the services you don't use.

As much as I enjoy the thought of a Netflix where I'd pay for how much I actually watched (like how cloud hosting is billed). I think it also opens the possibility that they'll be creating content that abuses that notion. Like, longer content for pay per minute. Or longer seasons for pay per episode.

But the idea sounds great to me.

Frankly Netflix scares me and I am scared more by the proliferation of the Netflix model to video games.

What I saw happen with television is that it went from a ratings based model (they had to make stuff people wanted to watch) to a model where the "cable bundle" was determined by cigar-chompers in a dark room somewhere. I can say I want this group of 20 channels, but I can't say I want CNBC but not CNN, Fox and MSNBC.

They get paid anyway so now there is no connection between "what I want as a consumer" and "what I get". So of course you get slow decline like we've seen with the cable industry -- the only meaningful conversation you can have through the market is "exit".

Netflix has only "exit" and "not-exit", it doesn't have market signals that say you can do 1% better and make 1% profits. So ultimately it gets a dull edge.

So many firms are falling over each other to offer you all the Madden NFL and Assassin's Creed you could possibly play and I am frustrated that people don't perceive that this has happened to MTV and most of the other cable channels since the 1980s: when I was a teen we watched music videos, but a 70-year old man bought the network and decided that he didn't want us to watch music videos. Once Youtube let people chose what to watch we found that people still love watching music videos: does anyone watch MTV? I'm sure there is somebody long past youth in Hollywood who thinks they can stay in touch with "youth culture" by watching MTV. I suspect there is a TV set in a Nielsen home that blasts away 24 hours a day with nobody watching. Other than that I don't know.

Music royalties and professional VJ's (read, unionized) eat into those profit margins. That 70 year old man figured out he could cut costs by hiring unprofessional "reality" actors and running his network on the sweat of young perma-lancers.[0]

[0]https://www.dailykos.com/stories/2020/8/12/1968730/-Sumner-R...

You’re not wrong, but it’s more complicated than that.

TV networks initially embraced reality television because of the Hollywood writers strike. The writers stopped writing, so the networks turned to content that didn’t require writers. Then they discovered that reality television is cheaper, too. The rest is history.

>when I was a teen we watched music videos, but a 70-year old man bought the network and decided that he didn't want us to watch music videos.

That's funny, but empirically that's not what happened. A quant working at MTV realized that reality TV like Real World, Jackass, and The Osbournes were way more lucrative and way more popular. At the same time music exploded in access (Napster). People voted with their eyeballs.

To make sense of the decisions that the TV industry makes you should also consider what their moat is and how they tried to protect it from the internet.

Studios still made content people wanted to watch, monetizing it became a lot harder which is why (strictly) ratings became less important. Bundling came around because it was easier for Disney to force Comcast to swallow channels like Lifetime and Toon Disney if they needed that to get access to ESPN. ESPN (live sports) has no digital compliment and for many people is the only reason they pay for Cable. The name of the game then became "How much money can we extract out of the remaining ESPN subscribers".

I think the last part is right. In reality it is slow motion suicide to lose engagement with your customer.

That is, the industry would have transitioned to a "harvesting" model with increasing cable bills and fleeing subscribers no matter what the competitive threat (internet, netflix, ...) is. The competitive threat determines where the crisis happens, but the crisis is built into the product.

I don't believe it about MTV. I don't think anybody watches the Real World, Jackass, or the Osbournes. They read in the newspaper that somebody got killed acting out a scene from Jackass or they watch the highlight reel for the Osbournes on Entertainment Tonight. The meaning of that kind of television is not in the watching, but in being a subject for the rest of the media.

What would be the difference between Siftrics vs standalone OCR software? Does it handle complex documents better or something?

>Does it handle complex documents better or something?

Exactly. It can handle tables of information with a variable number of rows from document to document. It can also handle arbitrary rotation, skew, offset and cropping. To see an example of some extreme cases: https://siftrics.com/hydra.html

>What would be the difference between Siftrics vs standalone OCR software?

The reason I started Siftrics is this:

Lots and lots of businesses need to put data from their documents into a database. Standalone text recognition gets you 75% there.

I thought, at the end of the day, companies still have to hire engineers to write programs --- which _do_ leverage standalone OCR/text recognition --- that are specially tailored to their documents.

I want to eliminate those specially-tailored programs. Now Hydra isn't perfect, but in many cases (people are willing to pay for it), Hydra reduces those specially-tailored programs to a single function call...

  client.recognize('avionics-invoice', ['invoice_1.pdf', 'invoice_2.pdf'])

...followed by your database insert. :)

Haha I've built a very rudimental version of this for my personal documents. No grouping or categorizing of text areas, just PDF->OCR->Database. I do scan for letterheads, though, and save them in a separate table.

A question, if I may: How big is your userbase?

We did the same, it's I think one of our best moves because everyone just gets the same price, nobody feels like they're missing out and it's super easy to administrate so we can concentrate on the product, not the pricing.

I have to say — nice product. It’s very clean. I can see how people are convinced to try it. I want to sign up just to experience more of the buttery-smooth UI.

I don't see much difference between your "products" and "plans with different feature sets", except yours is priced per use instead of as a subscription.

Also, 'First 1,000 pages free, then'. GP - I can put this into columnar format for you, if you want...

Is siftrics using or considering to use the current state of the art?https://paperswithcode.com/sota/optical-character-recognitio...

https://paperswithcode.com/task/scene-text-detection

Yes! We handle human handwriting and more than 100 languages.

I encourage you try it yourself. After all, you get 1,000 pages for free :)

wow, your product offering is great!

I could see it being very useful to small businesses that work with paper a lot.

A minor nit, perhaps, but you have a jarring typo on the home page: "software devleoper".

Reply on news.ycombinator.com