Is the reason so many companies fail to capitalize on data science and machine learning because of a lack in talent or because the business does not spend enough time understanding its value?

The major reason, in my view, is that business does not give it the due recognition and does not ‘wish’ to understand its own data

it starts with Analysts which are to “lazy” to write UDFs (use technology at all its usage/possibilities) and as mentioned think that they don’t need to know about every attribute in the tables of accessible DBs. Then data/accessibility is a huge issue and there we are on the responsibility of the management. Project manager without knowledge and interest in DS/ETL/data makes the projects fail. If you are not forcing any decisions to establish new data access, technologies and processes neither DS nor the company with get any bit ahead.

one of the main reason is the country that you are if there is not organization to record data managment e.g. patiant in and out this will be obsticale to bring data to analysze , if the government can funding to make this happned, this will make task so easy to open this industry and make it compatitive.

In my opinion
Talent is exist but companies do not understand the value
In my work little companies know the difference between visions and strategies
They analyse data to solve temporary problems
I think when we work in this job we got solutions for every thing in the end of work.

There are tools to understand and capitalize on data volume and real-time analysis capabilities, but it takes human knowledge and dedicated time of team’s to transform data (even if already in KPI’s) into real usage outputs (either conclusions, trends, profiling, and strategic data). So, yes time and usually human resources.

A lot of telcos have outsourced different parts of their network and IT infrastructure, which makes is practically difficult to pull together data. There is also still silo mentalities inside the operator which in turn prohibits cross dept collaboration. I would also want to add that the implementation of solutions over time can in itself make it very difficult to get the right data, clean it sufficiently well to get value out of it. Historically, and my final comment is that the graveyard of failed IT/OSS/BSS implementations is large, so a CFO/CEO would have the right to be skeptical about the ability to succeed.

You need to create transversal teams to deal with so innovative and complex projects. Try Agile team working (sqdas, etc…) . All the same be sure that data is really accessible; sometimes data systems are not prepared.Finally avoid proof of concepts: go directly to cases into production even if the scope is less ambitious.

In my opinion it’s mainly the following reasons:
- data stuck in legacy systems, often in silos
- to little management attention / willingness to invest
- real data experts rare
- lack of exiting and proven use cases

Thank you all. I also see that enterprises and their employees are not AI / ML ready.

Its a mix of things but in my experience boils down to companies not being clear on what data to use and for what purpose and not having or engaging the right staff or support to deliver a functional solution. It will come in time and in an incremental way rather than a big bang approach for most organisations.