With managerial changes including the departure of European boss Kasper Rorsted and global server executive Peter Blackmore, missed financial targets and a sharp fall in stock price, the new HP seems to be faltering.

The consensus is that HP initially did a tremendous job managing change after acquiring Compaq. However, the recent news underlines the difficulty of making sense of the multiple server platforms and organisational processes the combination entailed.

When HP chief Carly Fiorina met her Compaq equivalent Michael Capellas, both saw stars, but the baggage of the past has a habit of weighing down even the most promising relationship. And Compaq brought more baggage into the relationship than Elizabeth Taylor. Before buying Tandem and Digital, Compaq was a Wintel juggernaut; after buying them, it shouldered a broader load including OpenVMS, Tru64 Unix, NonStop, Mips and Alpha. Even before Compaq, HP already had PA-Risc, HP-UX and the ancient 3000 servers.

HP has recently been reorganising its operations to improve cross-selling capabilities, but cutting through this Gordian knot of chips and operating systems is HP's big issue. Such pre- and post-merger moves as canning Alpha, Tru64, PA-Risc and the e3000 caused ructions and perhaps chickens are coming home to roost. Certainly the server businesses of rivals such as IBM and EMC seem to be standing up better by comparison.

HP must also address the issue of its commitment to Intel's 64bit Itanium, the chip it helped develop and towards which it is pointing server customers. HP didn't quite bet the farm on Itanium but it came pretty close; little wonder that it has recently been busy de-risking through support for Intel and AMD 64bit-extended processors. A costly new processor with limited support for the installed base of software was never likely to be a quick sell but Itanium is slowly catching on.

Could HP have served its customers better? Perhaps. The server changes it has made beg comparison with IBM, which has spent more on software and services than hardware acquisitions, and has successfully begun fusing what were the RS/6000 and AS/400 servers on the Power architecture.

This month's announcements of Itanium and Alpha servers running the latest HP-UX will help, but the company will need to show customers just how much it cares when they come to the server crossroads. User group Interex suggests that firms are worried by a lack of direct support - a notable omission at a time when buyers making once-in-a-generation server changes will demand to feel love from vendors.

HP may have hit a speed-bump but it has opportunities everywhere. Its acquisition bid for Synstar suggests one obvious route - buying small services firms with regional strengths and domain expertise. Low-key acquisitions also suggest that HP is bolstering its position in systems management software, without succumbing to the temptation of challenging IBM in other enterprise software markets.

This is the way forward. HP has great R&D, a brand that is a byword for reliability, and a terrific customer roster. Being the all-round IT partner without software bias suits HP, but more direction and fewer vague references to the "adaptive enterprise" are needed.