Salesforce's $20 billion challenge: to go deep, deep, deep

Going vertical

But breadth isn’t sufficient for Salesforce, it needs depth – and herein lays a fundamental dilemma — something that’ll see Salesforce tread a tight rope.

On the one hand, Salesforce needs great verticialisation, essentially apps on the platform and part of the platform making more useful to customers in different sectors.

Going vertical is the thing that’s helped make Oracle and SAP indispensable to many, and broad. It’s something Salesforce is now engaged on, under Block.

Salesforce has relied heavily upon partners to deliver the apps for its platform but, to date, that’s not exactly been a runaway success.

The AppExchange market place, launched in September 2005 with 70 applications, and hit 1,996 apps by 2013, when the company announced it would roll white-label exchanges for customers. Block claimed the platform is now resident to 2,700 apps from ISVs and four million apps built by those customers. If Block is correct, then the number of commercial apps has grown by 704 since 2013.

Andrew Lawson, Salesforce senior vice president and UK leader, told The Reg more vertical apps are coming but warned it’s a long-haul: “This is Keith's baby and you will see us putting more and more investment into that ... It’s going to take some time.” Lawson, like Block is ex-Oracle.

“You will see some products coming from us in financial services market, things in retail, looking to work in partnerships as we are doing with Sage to develop the capability," he added.

But going vertical is risky: you need to know when to stop so the overlying platform remains application neutral and so you don’t start invading partners’ markets.

The other problem is being able to knit everything together on an API and at the data level. The risk here is that Salesforce becomes a consultant’s dream and a customer’s nightmare.

The pressure is there already.

The UK and EMEA vice president of Salesforce partner Cloud Sherpas, Colin Robinson, told The Reg his firm is moving into “wall-to-wall business transformations” that cover the full customer lifecycle from acquisition to support during and after the sale is made.

The question for Salesforce is how far it responds to this and tries to become the system of record for customers. That’s been the mantle of ERP, but Salesforce isn’t ERP.

Salesforce has already taken steps to cater to this: it spent $390m last year buying RelateIQ – an enterprise relationship manager that searches unstructured data in emails, social networks and calendars to track relationships among customers.

Darren Starr, chief executive of Salesforce consultant, StarrForce, explained to The Reg the kind of dynamic tension that Salesforce now faces. “People think a good CRM should be able to pull data from different systems so you aren’t constantly re-entering data – that’s the RelateIQ value proposition,” said Starr.

The question, therefore, becomes how deep Salesforce should go, and in what areas.

Take Configure, Price, Quote (CPQ). Salesforce has given choice, with Apttus, Steelbrick and Chameleon, but CPQ varies not just by business units, geographies and verticals, but also within the same vertical, such as finance.

Siebel – lost to Oracle — built the last-word in CPQ — but in so doing Siebel produced a monumental stack that became difficult to customise. It started as the alternative to Oracle and SAP (sound familiar) and it ended up being hated by customers.

Adding more to the Salesforce stack wouldn’t just make it harder to customize: it throws up integration challenges, across apps on and off the platform.

Integration is one of those how-long-is-a-piece-of-string questions in business. Salesforce already offers integration to enterprise cores SAP, Microsoft and Oracle via Lightning Connect, a framework, but the devil is in the detail.

But, notes Robinson: “You still need strong integration to the system of record, that’s ERP, and that’s a traditional IT challenge, not just for Salesforce.”

If you think ERP integration problems are a problem of SAP and Oracle, think again. Salesforce projects can, and do, already run aground.