Field notes. Every BPO leader ends up arguing about this: should this client get its own team, or should its work sit in a pool with other clients’ work? The argument is usually conducted as if one answer were right. It is not. Both models are good tools. Most failures come from picking the wrong one for the work, or from calling a shared pool “dedicated” in the sales deck and then raiding it when a bigger client has a bad day.

Definitions first

A dedicated model assigns named people, a team leader and often a floor area to one client’s process. Their time belongs to that client; when volume is low, they are idle or on training. A shared services model uses the same pool of trained agents across several clients or several work types, with work routed to whoever is available. The words are also used inside a single company: a “shared services centre” handles finance or HR work for many business units.

There is a middle path as well. A core dedicated team handles steady work, and a shared overflow pool absorbs peaks. This is what I recommend more often than not.

Why pooling saves money: a small example

Illustrative numbers. Client A is busiest in the morning, needing 10 agents until 1 pm and 6 after. Client B is a mirror image: 4 agents in the morning, 9 in the afternoon. Staffed separately, A needs 10 and B needs 9, a total of 19 agents, because each team must be sized for its own peak. Pooled, the morning demand is 10 plus 4, which is 14, and the afternoon is 6 plus 9, which is 15. The pool needs 15. That saves 4 agents, about 21 per cent. At ₹35,000 per agent per month in cost, the saving is ₹1,40,000 a month, and it grows with the number of clients whose peaks do not line up.

This is the same statistical effect that makes one large queue more efficient than several small ones in contact-centre theory. Larger pools smooth out randomness. But it only holds if agents can actually take both clients’ work, and that is where the costs of sharing appear.

What pooling costs

  • Skills. Every extra client process an agent must know raises training time and error rates. Cross-skilling two similar processes is manageable. Cross-skilling five different ones is rarely clean.
  • Security and compliance. Some clients require a ring-fenced area, restricted access to their systems, or no mixing with other clients’ data. Regulated work in banking and healthcare often comes with these conditions. Check the contract before you assume a pool is allowed.
  • Quality and brand voice. A dedicated team absorbs a client’s tone and product knowledge. A pool agent who handled a telecom query ten minutes ago and a bank query now will be a little less sharp on both.
  • Measurement. The client wants to know what it is paying for. In a pool, billing requires allocating time or transactions accurately; any sloppiness in that allocation becomes a dispute.
  • Management. Whose targets does the pool team leader serve when two clients’ queues are both red? Without a stated priority rule, the loudest client wins.

What dedicated costs

Idle capacity, mostly, and a higher cost to cover absence. If a dedicated team of eight has two people absent, the process feels it at once; a pool of eighty absorbs the same percentage with less drama. Dedicated teams also create awkward moments in the contract, because the client expects the named FTEs to be there and will notice when you borrow them.

Where dedicated wins clearly: complex or regulated work, long training time, a client that wants a visible, ring-fenced team, strategic accounts, and any process where product knowledge is the whole job. My opinion: if training takes more than a few weeks, think hard before sharing the people, because the saving from pooling is eaten by the quality loss.

Where shared wins

Standardised, rules-based work with short training: invoice processing, data entry, basic email handling, level-one queries on simple products. Also small clients whose volume would not fill a team of reasonable size, and work with predictable peaks that differ across clients. For internal shared services, standardisation is the entire benefit: one process, one tool, many users.

Rules that keep a pool honest

  • Write the priority rule between clients before the first peak, not during it.
  • Cap the number of processes per agent. Two or three is a sensible starting limit; test beyond that only with data.
  • Track quality by client within the pool, not only for the pool as a whole, because an average hides a client being served badly.
  • Be honest in the proposal about which model you are offering. Selling “dedicated” and delivering shared is a trust problem waiting for an audit.

One more point, which tends to be skipped. Models change as an account matures. A new client often starts dedicated, so the team learns the process properly. After a year, if volumes and procedures are stable, some of the work can move into a pool with the client’s agreement and a price reflecting the saving. Moving the other way, from shared to dedicated, happens when volume grows or quality complaints show up. Treat the choice as reviewable, with triggers you can name, not as a decision made once at the start.

Figures and names in the examples are illustrative.

Leave a Reply

Trending