In a research note, Sanford C. Bernstein analyst Toni Sacconaghi expressed doubts that IBM’s much-trumpeted transformation—which relies heavily on its cloud, mobility, analytics and security businesses— is getting traction, or at least enough traction to offset the declines in its legacy businesses. Sacconaghi, a respected figure among investors, echoed concerns others have about IBM (IBM) and which IBM execs have sought to downplay.

Sacconaghi’s bottom line:

The “theory” is that as the company’s strategic imperatives become larger over time, IBM’s growth rate will improve. The problem is that is hasn’t. While the strategic imperatives have grown $13.8B over the last 4 years (including acquisitions), IBM’s “core” revenues have declined by a stunning $29.7B, resulting in revenue declines (adjusted for currency and acquisitions and divestitures) in each of the last 4 years. Most sobering, IBM’s revenue growth rate on this normalized basis has not improved during the period.

The company continues to bet big on emerging markets including healthcare data analytics. Last week it announced its plan to acquire Truven Healthcare for $2.6 billion, for example. That’s a bold move, but skeptics wonder if these continuing investments will reap the necessary rewards. And there is some feeling that IBM is buying not only data and technology but customers for Watson, its marquee self-learning or artificial intelligence software, instead of selling that service to new customers the old-fashioned way.

For more on IBM and health data, watch:

A few weeks ago on IBM’s fourth-quarter earnings call, chief financial officer Martin Schroeter said the company’s performance in these strategic areas “accelerated” in the previous fiscal year.

“Together, cloud, analytics, mobile, social and security grew 26% and delivered $29 billion in revenue,” Schroeter noted. Together, he added, they represent more than one third of IBM’s business, up from 22% two years ago.

IBM, which had no comment for this story, has also claimed huge uptake of its Bluemix software development platform by thousands of new developers every week. But there is still a lack of big-name customers for cloud and data analytics services. Lots of partnerships and acquisitions for sure, but customers? Not so clear. Meanwhile Amazon (AMZN) and Microsoft (MSFT) continue their well-funded public cloud push.

Get Data Sheet, Fortune’s technology newsletter.

Sacconaghi’s report comes just days after Morgan Stanley (MS) analyst Katy Huberty raised her rating of IBM from “equal weight” (or neutral) to “overweight” (or buy.) Huberty said the company is doing better at this multi-year transition than is generally accepted.

Sacconaghi plainly disagrees. He said shareholders need more detail about IBM’s key businesses, including software. For example, how much has that business declined and how much of its revenue comes from software that runs on old-school Unix-based hardware? The hope is that these questions will be asked (and answered) on IBM’s analyst briefing day this Thursday.

And, in cloud where IBM competes with Amazon Web Services in shared public cloud infrastructure and with a number of software-as-a-service and platform-as-a-service players, a key question is how much of IBM’s “as-a-service” revenue comes from new customers and how much is due to a shift of existing customers from on-premises use of software to the newer model?

In Sacconaghi’s view, the jury is out even though IBM’s priority businesses are doing well.

The problem is that Big Blue’s overall growth is negative and not improving, he noted. That suggests, he wrote:

that IBM is losing wallet share from its customers, and market share of overall IT. Overall IT spend over the last few years has averaged about 2% – 3%. IBM’s revenues have been declining at about -2%. Moreover, IBM’s financials suggest that its strategic imperatives don’t appear to be improving its wallet share at customers – rather, since overall revenues are contracting, it suggests that IBM is not even keeping up with its customers’ shifting priorities.