Very neat, didn't know about Deming Regression. Hard to imagine a case now where it wouldn't be the more appropriate method.
1,082 karma · joined June 24, 2016
Imagine it's 2014 and you're deciding between two hot new framework ember and react, this is not just a question about what is hot or shiny and new.
Instead of:
"Benchmarking Multimodal RAG through a Chart-based Document Question-Answering Generation Framework
Multimodal Retrieval-Augmented Generation (MRAG) enhances reasoning capabilities by integrating external knowledge. However, existing benchmarks primarily focus on simple image-text interactions, overlooking complex visual formats like charts that are prevalent in real-world applications. In this work, we introduce a novel task, Chart-based MRAG, to address this limitation. To semi-automatically generate high-quality evaluation samples, we propose CHARt-based document question-answering GEneration (CHARGE), a framework that produces evaluation data through structured keypoint extraction, crossmodal verification, and keypoint-based generation. By combining CHARGE with expert validation, we construct Chart-MRAG Bench, a comprehensive benchmark for chart-based MRAG evaluation, featuring 4,738 question-answering pairs across 8 domains from real-world documents. Our evaluation reveals three critical limitations in current approaches: (1) unified multimodal embedding retrieval methods struggles in chart-based scenarios, (2) even with ground-truth retrieval, state-of-the-art MLLMs achieve only 58.19% Correctness and 73.87% Coverage scores, and (3) MLLMs demonstrate consistent text-over-visual modality bias during Chart-based MRAG reasoning. The CHARGE and Chart-MRAG Bench are released at https://github.com/Nomothings/CHARGE.git."
Give me: " "AI Fails Hard at Reading Charts: New Study Exposes Shocking Weaknesses!"
A groundbreaking study reveals that even the smartest AI models struggle with charts, scoring just 58% accuracy—proving your brain might still be better than AI at decoding data! "
I agree the reading time does not really convey too much useful info, the blogging platform is designed for simpler and shorter posts. Maybe a replacing it with a wordcount would be better.
If there are any points I can clarify further let me know, thi s is a really off-the-cuff sketch and I had fun doing it and wouldn't mind digging a little deeper on some ares.
I am currently a developer in charge of growth at photoeditorsdk.com and about to crank on our inbound marketing. Although a very different market from your's we've had success with cold-emailing as detailed in [Predictable Revenue](https://www.amazon.com/Predictable-Revenue-Business-Practice...). The biggest boost came from preparing industry specific sales pitches & demos.
Cheers
In general, what stands out to me is that you do not have a clear USP and you do not seem to speak your target markets language, it might be that since you are addressing markets in many different verticals that you have no clear persona in mind to convince with your copy, no specific set of pain points or benefits to address directly. In any case, I would try to interview my current customers, ask them to describe in their own words how your software is helping them, what pain they were experiencing before and how they previously dealt with it. Then you can go about ‘cloning’ them by marketing to specific verticals, have landing pages for hospitals and retailers, agencies and address problems specific to these industries using the information and sound bites you mined from your customer research. This will also allow you to do some long-tail SEO, produce relevant white papers, tailor your testimonials and in general tighten your copy since you now know who you are talking to.
Take a look for instance at your benefits, they are not benefits, at best they are product features: “Reminders & Communication”, and some are not even that: “No Install Or Setup”. How about “Never fail audit xyz again with our bullet employee foozle tracking system”, find a specific problem and tie it to what your software can do. There is an interesting line on your landing page: “Stay on top of critical legislative and legal requirements to store and recall employee.“ How many businesses are not compliant with these requirements, what are the consequences, use this risk and fear in your marketing, it is a strong motivator and sure to catch the attention of any business owner or HR manager, HR is inherently boring if you can’t excite at least agitate.
This is also a great opportunity to educate, small business owner, entrepreneurs, agencies, we all like to skimp on HR, do more research on the legal risk and host a webinar for business owners educating them on it, underscore the necessity of reliable HR management for becoming a ‘serious’ business.
I also found there was a lack of strong social proof on your site, this is HR show some people, preferably happy teams, show company logos, media mentions, photos of the people you quote. If you have impressive sounding metrics such as the number of employees processed by your system, total payroll handled etc. display them somewhere.
Also, if you aren’t already, try outbound sales, email people, get them on the phone, discuss their problems and whether there is a mutual fit, maybe you can uncover a great selling point, niche or approach, read Spin Selling and Predictable Revenue. If you can get this channel to work profitably and in a predictable fashion, you have total control over your revenue. If you not going full outbound make it easier to contact you, many companies, especially old economy want to talk to someone on the phone, no-touch sales on the internet is a tough proposition in some industries.
As to pricing here you probably have the easiest wins, get rid of all your plans (especially the free one) charge 19 bucks or something per employee in the system, this is a true value metric, easy to understand and uniform. Right now you have a large number of fairly arbitrarily split pricing plans, with non linearly increasing prices, why do bigger companies pay less per employee, if anything you should find a way to have them pay more. Not only will this increase your revenue, you will also grow linearly with your customers. Once you have built out quantitative buying personas, that is quantifiable sets of customers that differ in they price sensitivity, budget, the features they value etc. (most likely corresponding to the different verticals you serve) then go about and segment. Have tiers actually cater to distinct customer groups segmented by meaningful criteria such as specific feature sets, integrations with industry specific software etc. see here http://www.priceintelligently.com/blog/everyone-in-saas-is-u... for more info.
You also mention HR consultancies, is there a way to partner with them, offering either a white label version of you software they can sell to their clients or a commission, are there opportunities for other types of partnerships; events, conferences, case studies etc? You seem to have a great product in a good market I hope you have commensurate success :)