There are a few questions in this discovery journey of mine about the stickiness of client to a data center facility.
Some data center service providers give too much credit to the “stickiness” of client or give it as an easy-to-understand-answer to the media when asked if they are confident of retaining their existing client portfolio.
Some data center service providers are enlightened to regards client stickiness is not so apparent as before and are doing some major mind set shift and enhancing their solutions and their interactions with their clients.
Below is my own take on this topic.
What is customer stickiness?
Marketing books has defined this as customer loyalty, which is associated with brand.
I like Wharton business school’s definition of customer stickiness which is the increased chance to utilize the same product or service that was bought in the last time period (see reference 1).
In my opinion, in the data center industry, brand does not matter much when it comes to co-locate or to stay co-located in a particular data center facility.
I think it all boils down to three main factors:
- perceived value (see reference 2 and 3),
- risks; and
- most important of all, cost.
In the first place, when customer come to co-locate with Site-A, it has gone through rounds of evaluations and decided Site-A delivers the best value. And value, doesn’t mean low price. It means all things considered and after trying to normalize all the submissions to the RFQs, a particular vendor offer the best combination in terms of price, features, carriers connectivity, services, etc.
After they have chosen a site as their co-location service provider, let’s just called Site-A, and the honeymoon begins and perceived by Site-A to last forever. That at least is in the perception of the Site-A management.
In another industry, the cut throat world of airline business, customer vote with their feet unless there is limited choice for getting from point X to point Y and they are resigned to the one airline that serves that route. Site-A management may think that their client has limited choice because the client is bounded by value, risks and cost combination.
Data center co-location is a cost. Staying put or moving away, both carries a cost. This cost should be further broken down into initial setup cost, migration cost and operating cost. There are also risk factors to consider. Such costs and risks are considered for the two choices:
- Staying put where they are, or expand at where they currently with.
- Move to a new site
If you are a data center operator with healthy weighted average leased expiry (WALE) and that is your main focus, you may be thinking that customers are sticky and are in no danger of losing that account. You are wrong in a big way.
What are the reasons for the “stickiness” of existing customer to a data center facility?
Going back to the Wharton definition, what will increase chance for customer to stick with you and renew their co-location contract?
- Risks, especially it is mission critical to the client’s business
- Compared to moving away, the initial cost is higher than to stay put (even after lease renewal increment)
- Migration cost
- Shortfall of service from alternative data center service providers
Some organization, have planned for data center renewal, i.e. they review the relevance of the data center facility where they have colocation or private suite to decide whether to stay or move. I know of one global bank that will refresh one data center per region on an annual basis.
These days, companies frequently go through M&A, and their data center may be closed or the data/applications moved elsewhere on a short notice.
How does new data center business come about? At times it is due to their current data center space cannot serve their current or future needs.
And client suddenly terminating and moving their site, though rare, is not unheard of. I have heard of a extreme case of sudden termination of a major contract with a site in Asia due to continual displeasure by the client of the co-location service availability and multiple response time slippages and the client just terminate and pay the penalty. That is one pissed customer and the gain to another data center. We cannot afford to take things for granted.
Let’s examine what makes data center service provider thinks that their clients are sticky:
- The setup cost
- Cost of migration
- Risks of prolonged downtime
- Marginal gain (if any) of lower TCOs weigh against the above factors
- Facility people, whom the data center service providers are used to negotiate contract with, are more happy to stay put
BUT, the client may see a totally different picture:
- The current service provider service or the site cannot meet current or near future needs (e.g. power capacity per rack, less service offerings than one that can provide cloud, network, etc)
- Their perspective may not even be a move, it’s decommissioning while commissioning a new site
- Even for a move, they have lower cost of migration compared to the old days of portfolio of critical IT applications that nowadays are split into multiple smaller application systems
- Risks of prolonged downtime can be managed through migrating an application at a time, development and testing first before moving the production environment, or other proven migration methods.
- TCOs savings, e.g. do you have continuous energy efficiency effort, and do you reduce your PUE factor as part of utility cost recovery?
- Enabling new capability through putting more of their required IT capacity (development, test, production) onto cloud
- Merger and acquisition, or a strategy to rationalize the data center capacity that they need
- IT calls the shots. Facility people in enterprises these days now listens to, if not already reporting to, the CIO and his team of IT project managers
Initial cost and migration cost are a major concern, but so is staying relevant in the business competitive world which places more demand on IT being agile and flexible.
And organizations have considered or are already moving more applications, some of it mission critical, to the cloud. (see Forbes article in one of the reference link).
How about challenges brought about by technological change? Active-Active data center set-up, Active-Backup data center set-up, has enabled client to plan for migration more easily than ever. And virtualization has made IT servers a commodity, meaning the IT resources are moved virtually across data centers which can reduce the risk of downtime than possible previously.
What are causing the decrease of stickiness?
Reason number 1 is demand by the top management for IT to be agile and support or even drives the business. This has to be a top-down total solution. The client will engineer, get consultation, change development and production strategy, use technological tools and capacity (e.g. cloud) to have that agility.
A lot of organizations have gone asset-light. Banks, financial institutions, big manufacturers, big service organizations, large retailers, have gone the route of leasing their IT needs from IT companies and hosting them in third party co-location service providers.
The nature of the IT applications development methodology of prototyping based development and rapid application development, DevOps where developers and Operations making software development, testing, and release more reliable and faster. This creates a demand that the testing and development of a data center space will not need to be higher tier, while demand for agility may mean a production environment will sit on a virtualized IT resource pools that can more easily sit in multiple data centers or move to a new virtualized environment more easily than previously possible.
Moving an entire IT development, testing and production environment is no longer a daunting project except for a small proportion of enterprises that have got mission critical applications tied to a monolithic system, but that has become more and more rare. CIO will plan to make himself a part of the business solution rather than problem which means IT agility is the norm these days.
Together with the following IT tools and new way of getting IT performance and capacity:
- commercial off the shelf software (COTS) instead of bespoke software is the norm
- Cloud (public, hybrid, private)
- Multi-tier and distributed application architecture
- Leased IT hardware and leased IT software
Public cloud and private cloud enabled through VPN over MAN/WAN has enabled company to use bare IT resources outside of their current network boundary, and DCaaS will breakdown the physical boundary of a data center facility.
Some organizations are doing away with physical data center co-location altogether. Ray White, a 100plus years old real estate sales and rental organization has moved its applications onto Google Apps and email.
Active-active production sites are more the new norms. So is it that hard to really perceive that your client is not that sticky anymore?
What can you do to increase the “stickiness”?
Firstly, shift to see the data center facility through the eyes of the clients, i.e. what is the perceived value? If it is just brick and mortar plus conditioned back-up power and controlled temperature environment, then what you have is cost differential which can be easily replaced by new entrants or someone with better economy of scale.
Fortunately, many data center service providers have begun to look into the value question and some are addressing it through going up the value chain. Such as DRT buying up Telx, most if not all are looking at reducing the opex by looking more seriously into the energy efficiency and helping clients saves money. Co-location service providers buying up other co-location service providers is a way to add scale and reach and perhaps drives down cost to be more cost competitive, however this is just my speculations. More likely, co-location service providers are now offering network services, a connectivity hub and inter-connecting service for within and across their key network hubs, and setting up their own cloud offering or partnering many cloud service providers.
The key here is, talk to your clients, before they called a RFQ/RFP which will be too late.
So, look at those factors:
- Client considering to commission a new site
- TCOs savings
- Enabling new capability through putting more of their required IT capacity (development, test, production) onto cloud
- Newer, faster deployment, modular build, and have IT/networking capability in-house to talk to the client IT people
Going back to the airline industry example, the full service players in the airline industry are not sitting idly while new entrants can drive value and have lower cost with the newer more economical fleet and tighter seating arrangements. The airplanes are the same product albeit with different configurations but it is limited. They distinguishes by the service that they give. They continue to innovate and reach out to its customers to protect their margin and ultimately their business. It is worth mentioning that in an industry where the average training duration is six weeks, Singapore Airlines flight attendants endure five months of schooling. Emirates Airways probably is close to the same.
It is no longer ok just to make your power and cooling infrastructure rock solid. It is to also be agile and akin to client’s IT needs. It is to be more relevant, and creates more services that the clients will use and increase that stickiness, if you will.
Make yourself more relevant to today’s data center clients.
It may well be that the cloud service providers and server manufacturers (in China, Inspur builds and own data centers for BAT type of clients, like what they already did in ChongQing and will be building in Henan) that will occupy lots of data center space as they in turn provides IT capacity and capabilities to the enterprises.
In summary, client stickiness in the data center industry is in my opinion, a myth.