Have we reached peak SI

Have we hit the peak for systems integrators (SIs) (just as we appear to have reached “peak oil”), and it’s all downhill from here? While SIs are doing well at the moment, structural changes in the IT market suggest that the long term forecast is not all sunshine and roses as some pundits are predicting. With IT spend migrating from IT departments (the SI’s traditional buyer) into the lines of business, the ongoing shift to smaller projects delivering on-demand (rather than on-premesis) solutions, and the replacement of traditional support arrangements with outsourced and managed services, it’s hard to see how SIs will continue to grow when demand for their services seems to be tipping into decline. Globalisation, software as a service (SaaS) and cloud computer are reconfiguring the IT landscape and SIs look like they will be the big losers.

Predictions for the continued growth of the SI market are based on the understanding that companies are consuming more IT today than they were yesterday, and the assumption that increased IT consumption will result in tidy profits for SIs. Predictions are a funny things though, based as they are on historical trends. Guess that the market will continue to rise when you’re in the midst of a bull market and you’ll be right, most of the time. That is until something happens, something you didn’t anticipate, something that catches you unawares. The assumption that SI revenue is tied to IT consumption might no longer be true. New tools such as SaaS and cloud computing are enabling line-of-business leaders to step around the traditional IT department and engage with technology directly, bypassing the SIs traditional relationships in IT and providing them with fewer opportunities to sell their wares. At the same time the shift from on-premises to on-demand solutions – solutions which the business is happy to rent rather than own – is slashing the effort required to install, configure and integrate these new solutions, often by as much as seventy to eighty percent. On-demand solutions also have much lighter support needs relying on self-help wikis, users forums and power users, leaving the SI with little more than a small help desk to manage. With only limited access to this new class of IT buyer, dramatically smaller projects, and lower support revenue, the SIs role as IT enabler seems to be in decline. All good things come to an end though, and you usually only realise that the end has come after it has already passed. IT consumption might be going up, but there’s a good chance that SI revenue could soon be going down at the same time.

SIs are fundamentally sandwich shops{{1}}. When we don’t have the time or money to maintain our own kitchen or make our own sandwiches it can be more efficient to head over to the local sandwich shop to pick up what we need. Their margins are thin and revenue is largely tied to the size of the sandwich you just bought, so they’d really like you to buy a larger and more expensive sandwich. (Notice how sandwiches have grown so much bigger over the years, and everything is now gourmet?) And, of course, pre-made sandwiches are always a lot cheaper than special orders. This sandwich shop model is something that was established early on in the history of business IT. How else could companies afford to access the rare (and expensive) IT skills they needed to create all the systems they need? This might be a payroll system, or stock management, sales pipeline reporting, or the dreaded enterprise resource planning (ERP). Consuming IT used to mean hiring an SI to build and integrate something for you.

[[1]]Business models for the old rules of IT @ PEG[[1]]

The world has changed since then. Back when I started in the industry my home computer couldn’t hold a candle to the beast I was given at work. Today, however, my shiny new 17″ MacBook Pro makes the locked down Windows XP laptops I’m offered seem like a bit of a joke. A new breed of business manager has crept into the business while the world has changed, these are people who grew up with technology and are comfortable solving their technology problems on their own. They know that there are alternatives to the expensive solutions proposed by the IT department (solutions that IT will engage an SI to deliver), and they’re happy to use these alternatives. Why spend a seven figure sum and wait a year for the IT department’s perfect, enterprise-wide project portfolio management solution when there’s one that is good enough, one you can buy on-demand via a company credit card, and one which you know will be up and running in a couple of weeks? We might argue about the regret cost{{2}}, but the art of business is to make a timely decision and then make it work; it’s not to sit on your hands and wait for the perfect solution which will be delivered sometime in the distant future. While demand for new IT solutions might be growing, every time a business manager steps around IT to engage and on-demand solution SIs have one less opportunity to sell their wares.

[[2]]The price of regret @ PEG[[2]]

At the same time we find that these on-demand solutions – when SIs do get their hands on them – only provide a fraction of the revenue that a tradition on-premisis solution does. Time is money for an SI (literally, as most avoid risk by sticking to time and materials contracts) and fielding a SaaS solution takes only a fraction of the time required for a more traditional solution. There’s no hardware to commission with SaaS or cloud computing, nor are there disks to wait for or backup strategies to create. (You still need to worry about business continuity, but that’s another post.) There’s also little chance for customisation, and integration tends to be via standard APIs or pre-built adaptors. It’s not uncommon for a SaaS project to be eighty percent smaller than the more traditional solution. Fewer resources on ground and fewer billable hours means that that the SI can expect their revenues to head in the same direction: south.

We’re also seeing the erosion of SI support revenues. Support used to encompass both the application – in terms of application maintenance, patching and security – and the users – with training and a help desk. Many SaaS and cloud providers don’t want to provide traditional support services as it erodes their margins, margins based on huge scale and little human contact. One solution is to engage an SI to provide these services for them, either on a client-by-client bases or as part of some sort of alliance. A more attractive solution is to move – as much as possible – to a self support model where clients support each other via user forums or a Google search. We soon find that a much smaller help desk will suffice as it’s only required to be the point of last resort, or to support the more technologically illiterate users.

Taken together, these trends – reduced access to buyers, lower project revenues, and lower support revenues – seem to show that the future is not as rosy for the SIs as we first thought. Demand for IT might be growing, but growing demand for IT no longer implies growing demand for the services provided by SIs. The final nail in the coffin is the fairly recent move into SaaS by established IT application vendors. Microsoft has gone on record as wanting to capture a greater percentage of IT spend as license revenues, converting SI installation and customisation costs into licenses by providing clients with prepackaged configurations which can be turned on at the flick of a switch. Rather than pay for a SaaS CRM and then engaging an SI to configure it to your liking, you pay for the SaaS CRM along with a canned sales methodology (Miller Heiman{{3}}? Holden{{4}}?) which works out of the box (as it were). Integration between SaaS solutions is also being converted into a configuration option as SaaS vendors sign alliances – just as Google and Saleforce.com did with GoogleForce – enabling these alliances to offer complete application suites which work together out of the box.

[[3]]Miller Heiman: The Sales Performance Company[[3]]
[[4]]Holden International: Outsell You Competition[[4]]

Whichever way you look at it, now is not a good time to be a SI.