Hello Knime authority

  1. Is there any way to make a package for deployment using deafferent computation quota like data lore, those are highly affordable for individual data science practitioner?

Data lore (Jetbrain) offer 20$/Month for 750 hours of computing (Including all package and report generation support), why knime cannot offer that affordable for online deployment and may I know the explanation.?

I think it is mainly because they have different market positioning, business models, customer groups and development costs. The scale of the two companies is also different. Jetbrains has about 10 times the current employees of KNIME. You can imagine the difference in the number of customer groups they serve.

1 Like

@knime200087 you might want to take a look at the new options to host and run your knime workflows on the community hub in a dedicated team space.

Reading your post @knime200087 I feel it really well resonates with what I highlighted out as well:

In essence, an intermediate tier of the Hub for individuals and maybe with the option to only utilize the local interactive view but online to share with clients. Sharing reports with compute capabilities is called “Data App” and is only included in the Basic tier starting at 35.000 Euro.

Also worth pointing out that the compute costs, compared to an AWS Workspace where I ran for a former employer workflows remotely in batch mode, is substantially more affordable and comes at a fixed instead if flexible rate (because of the compute utilization credits).

Could you clarify whether the price is 35 or 35,000 ? If it’s 35,000 , then it might not be cost-effective for our discussion.

Well, there are two options. You can execute a workflow in the Hub with the 2nd Teams tier starting at 99 Euro / month plus the surcharge of the credits .

In my referenced post I did some math about a scenario I had and, for a modest execution of five minutes and five times each day that would incur an additional 75 € on the lowest tier. Not quite competitive compared to an AWS Workspace based on my experience but the Teams tier has some advantages in regards to collaboration.

Though, the really interesting things begin when you can actually share the insights and create wisdom for clients by sharing the results. For that to happen, as far as I understand the descriptions, the Basic Tier starting at 35.000 Euro is required:

About Data Lor from Jetbrains. I had a look and tried to compare their with the offering from Knime which is quite difficult. Truth is, the feature " Create BI apps with a few clicks" from Jetbrainsis already included in the 20 Euro package. Just by that, what Knime offers, is unattractive in terms for providing the results to clients.

Other aspects such as compute power is hidden by both vendors deep in their FAQs. So you have to be very cautious. Here is one example about the included compute minutes. Worth to note that I could not find any details from Jetbrains about the CPUs and RAM.

Choose Your Datalore plan | Datalore!

As it stands, to kick off working as a team or individual / freelance, I find it particularly difficult to find a compelling offer that would enable me to do some “data magic” and share the results with clients to build a business.

1 Like

Many people believe professional software should be expensive, but they may not realize how efficiently some tools can get the job done. For example, I use Knime for complex data cleaning tasks because it is versatile and powerful. However, when presenting to stakeholders, I often use DataLore instead since they are more familiar with python. My company is unaware that I initially use Knime for many tasks.

I believe Knime’s management should consider closing the free distribution of their base package and monetize by selling an offline version instead. They could still offer a limited edition for the open-source community. Additionally, they should make their data apps more affordable for mass adoption. Selling a high volume at lower prices can reduce production costs and increase profit margins, ultimately benefiting a larger portion of the data science community.

Currently, Knime competes with Alteryx, but given its versatility and potential, I think it should aim to compete with the flexibility that the Python programming language offers.

Many people believe professional software should be expensive,

Truth! I frequently face this as well. Maybe “you get what you are paying for” has some psychological effects. Though, from my perspective everything is a tool and it 99.99% depends on how well one can use it. I.e., what can you do with a barn full tools if you don’t know which and when to use it.

closing the free distribution
I don’t feel that way. That would very likely suffocate the community. Though, I must admit that I see other / additional monetization options to ensure development can excel like:

  1. More affordable Hub Tiers allowing to share interactive views (reports), no data processing, with clients using authentication
  2. Knime Professionals Network: Similar to Fiverr. A place where Knime Experts can offer their services. For each procured service, Knime could charge a fee.
  3. Knime Professionals Hub: Workflows by professionals, run on the Hub, on demand i.e. to crawl data from APIs or portals. Basically a hosted versions of the many awesome solutions found but companies would not have to setup things locally but could, on demand, get results very fast.
  4. Direct Workflow Support

Currently, Knime competes with Alteryx
That is exactly the tool a company I support made the choice, which I am really sad about, to switch to.

1 Like

Creative Ideas!!!

Also found reasonable to make “knime” think on their monetization approach! I know that, without data preprocessing it should and possible to make interactive report service to be highly affordable like “JMP Live” or Tableau Desktop subscription etc.


At least knime can think on new dimension on reporting, so that interactive doesn’t mean to be always on input dependent and of course without recalculation.


Hi @mlauber71, @takbb and @ipazin,

each of you being leaders in their regards I wonder if any of you, and apologize if I force loop you in but I attempt to keep this conversation alive since I personally feel it could boost Knime to new heights, have an opinion about the void in the Hub tier lists as explained above too.

Best
Mike

Hi @mwiegand , well the one thing I would immediately have a very strong opinion about would be the notion of “closing the free distribution”.

If I’m reading that suggestion correctly (and I know it was just a comment in the earlier post rather than an official position, before I panic anybody!), it would mean no longer offering KNIME AP for free, although maybe having a cut down community edition. I agree that this would likely suffocate the community.

Safe to say if it had been that way in March 2021 when I “discovered” KNIME, I wouldn’t be using it now and so neither would the guys I now work with who I recommended it to.

I had a discussion with one of the Alteryx management team at the “Alteryx Inspire” event in London, circa 2019, when I brought up the lack of a free/low cost offering of their desktop “designer” software. I explained I need something I can afford “between contracts”,so I’m keeping myself up to date, and I don’t mean a horribly cut down version! That clearly fell on deaf ears, and when I was next “between contracts” I found an alternative product, and haven’t touched Alteryx since.

To me the model of providing a desktop tool which everybody can afford, and then providing an enterprise solution on top is perfect because the software’s reach expands through word of mouth of the workers who need it and then the big bosses, who think that they need to spend money for something to be good, get what they want too. ( I’ve noticed they never want to spend their money in my direction, what does that tell me? lol).

In terms of hub tiers, yes I would like to see some additional tiers aimed more at individuals but offering some of the higher end features.

I find it frustrating that I can never gain first hand experience of the features of the higher tiers unless I I’m working for a company that has purchased it, and the only way a company I am doing work for is likely to purchase it, is if I recommend it… Which I’ll never do because I don’t recommend something I don’t have personal experience of. Chicken and egg.

If there were some entry level (restricted licence?) tiers that offered higher features for individual consultants to use indefinitely but with, say, limited “credits per month/pay as you go”, then I would personally feel I would be in a better position to recommend the equivalent (higher) tiers to my clients.

3 Likes

Wow @takbb… Crazy how exactly that post mirrors my situation and sentiments as well.

I have been trying to convince the KNIME team to change their marketing / language / offerings to a focus on affordable individual licenses, and less on “enterprise” solutions. I believe that their greatest growth opportunity is to make inroads into business users in small and medium sized companies. Having an open source desktop platform that is capable of nearly any automation, data processing and reporting task offers them a serious strategic advantage.

I would shift marketing toward “the power of one”, instead of only focusing on large enterprise / technical data science users. Stories that highlight the massive impact that a single KNIME business user had on an organization. Focus on organic growth of individual open source business users, and build a sustainable revenue model of affordable individual focused paid software and hub levels.

I will say that it will be challenging for KNIME to offer a low cost dashboard sharing solution that is competitive with a purely visualization presentation platform because of the nature of KNIME being a wide open blank building slate. How exactly would you enforce dashboards not performing extensive calculations? I know that I have never built a dashboard that didn’t have some sort of interactive calculation process involved. How exactly would you restrict the amount of calculation and processing? I think that is why they have fallen into a paying for processing time model instead of a low flat rate…

I would also like to see a simple low cost desktop add-on of a built in workflow scheduler for KNIME AP. Seems like a no brainer for the 1st step paid license option. Executing in batch mode via command line is far from user friendly for non techies… :robot:

1 Like

@takbb The question you mentioned is indeed a dilemma. To be honest, even if a restricted license is provided, it is actually difficult to use, mainly because of the complexity of k8s. But I think your idea is great, because after all, there are some avid KNIME fans, like you, who will study this thing thoroughly :slight_smile:

In addition to this restricted license, if it is purely related to KNIME, then I believe that hub (maybe in the future) can provide us with this restriction, but can experience and be familiar with everything.

In fact, I think it is like the KNIME Hub hacking activity conducted at the beginning of the year (testing the Business hub). It is also a good solution to provide a test hub and provide limited use. This kind of test hub does not need to be provided separately (after all, cloud services also require costs), it only needs to share a service with the KNIME test team (but the things on it may be reset or destroyed at any time, but it does not matter) , so that we can not only experience and feedback the latest features but also avoid the complexity of k8s.

1 Like

Wow, I never thought such an elaborate and swift feedback. Thanks all and it seems we all to a greater extend strongly believe this is a huge opportunity.

@takbb

If I’m reading that suggestion correctly (and I know it was just a comment in the earlier post rather than an official position, before I panic anybody!), it would mean no longer offering KNIME AP for free, although maybe having a cut down community edition. I agree that this would likely suffocate the community.

I didn’t meant or interpreted it that way to not offer a free tier. More closing the gab about “design” or report functionality which you also noticed being a huge opportunity.

I find it frustrating that I can never gain first hand experience of the features of the higher tiers unless I I’m working for a company that has purchased it,

Same for me. The way “distribution” currently works chokes of innovation as that pretty much always starts with the like of us having the freedom to roam wild.

@iCFO

I will say that it will be challenging for KNIME to offer a low cost dashboard sharing solution that is competitive with a purely visualization presentation platform because of the nature of KNIME being a wide open blank building slate. How exactly would you enforce dashboards not performing extensive calculations? I know that I have never built a dashboard that didn’t have some sort of interactive calculation process involved. How exactly would you restrict the amount of calculation and processing? I think that is why they have fallen into a paying for processing time model instead of a low flat rate…

My idea which I suggested was to only upload the final report on the corresponding port and combining that with simple server side checks about filters or even clear vCPU limitations when it comes to computation power. All the pieces are in place and my gut feeling tells me that this might be not too far off as there are even more complex features already in place.

@HaveF to you point about the costs. I’ve been running Knime in batch mode in the cloud on several servers instance classes and even the lowest tiers, if Knime is configured properly and isn’t fed humongous data sets, can run very well.

But that idea of a “Testign Hub” with no guarantee of data persistence is actually really great! It might represent a well balanced and almost immediate to archive approach for the Knime team to grant users, after applying for access, the ability to accomplish certain tasks. Though, that might be possible to get exploited easily using semi automation to force upload workflows accomplishing semi reliability.

1 Like

I was thinking the other day that it is likely possible to use KNIME AP to convert a basic program dependent dashboard into a standalone portable / embed-able E-chart dashboard. There are plenty of software companies that create embed-able JavaScript dashboards which could be used for code examples. You would need to include the underlying dataset and point the charts / filters toward them. It also seems like Chat GPT could be used in the conversion if it were provided a clear roadmap with code examples. I think you would likely need to pre-summarize data tables per chart to minimize the weight of the dashboard for better performance.

I haven’t hassled with the idea, because I can do so much more with interactive and flexible portable dashboards in Excel using the write to Excel Template approach.

@iCFO thinking out loud … with the constant change in the view nodes, where the most recent one uses Apache ECharts, I agree that what you pointed out comes into technical reach. Even Knime itself continues to shift towards an html-driven interface so in the greater scheme, a non-standalone, maybe even fully remotely browser app is what the Knime team aims for in the long term.

About what you wrote:

I haven’t hassled with the idea, because I can do so much more with interactive and flexible portable dashboards in Excel using the write to Excel Template approach.

I don’t like to use yet again MS Office or, for that matter, any alternative but feel pushed towards some hybrid approach given the constant change and reporting gap we exchange about. Actually, it makes me sad to say that since I have fallen madly in love with Knime back in 2016. Though, it not always can go uphill so I sincerely hope the team hears our voices :wink:

2 Likes

I was initially pretty disappointed when I had to fall back to Excel as my default for creating and disseminating interactive dashboards when clients didn’t have or want Tableau or Power BI. However, I have found that most people prefer the flexibility of a dynamically referenced Excel dashboard because of the endless customization options and the ability for users to do their own quick ad-hock side calculations.

I can also import the user’s entry back into my dataset for things like budget creation / modification. I now often embed Excel controller worksheets to manage user friendly interactive component settings in KNIME as well. The real time processing and visualization capabilities of Excel really fill a ton of gaps on the KNIME AP side, and KNIME does the heavy lifting which keeps excel files lightweight and responsive.

1 Like

Wonderful, you just made me realize “we” have gone full circle. Everyone hates Excel i.e. because of:

But since everyone uses it, everyone is obliged to continue to use it:

I just recently watched a very good video about why Raspberry Pi is so well established. In essence it comes down to a golden triangle of Performance, Support and Costs.

Excel pretty much nailed all and build upon its “legacy”. Contrary, Knime has a relative high adoption costs, because time. Performance for real workloads far exceeds Excel and the support, given this thriving community, is also superior.

But when it comes to the bells and whistles in regards to reporting … the cost of adaptation become excruciatingly high.

So, unless it becomes as easy to create a pie chart or better a report which tells a story and thus sells the idea, and I am explicitly excluding using AI everywhere because we all have (still) brains, Excel, tableau, power bi & co. will reign supreme.

We see the prove if that in the fact that even pros like you had to resort Excel despite having so much better options at your disposal.

1 Like

I spent hundreds of hours designing and over building a Tableau Write Back based universal visualization based granular budgeting and reporting software platform. It had some amazing drill down graphs, recorded user interactions and gave immediate calculation and visual feedback on changes. It also required a hundred workarounds to get around all of Tableau’s limitations.

To do it right at that point, I needed to combine Tableau, Tableau Write Back and KNIME Business Hub. (Although now I believe that I could likely patch a less attractive version together using just KNIME Business Hub alone now) Either way, it was a massively expensive / time consuming / relatively inflexible server based endeavor. I can now custom design something for each business that is infinitely more flexible using KNIME and Excel in no time. No struggles with Dynamic Chart titles, client customizations, filter value refreshing, write back, manual approval processes, etc.

3 Likes