Customer Experience
Two vendors quote the same seats, the same channels, the same hours, and the proposals land within a few percent of each other. One of them is selling you people who will know your product by month two. The other is selling you a place in a queue. The staffing model underneath is what separates them, and it is rarely the thing on the front page of the deck.
Nobody pitches you a shared pool. They pitch coverage, channels, service levels, a number of seats. The staffing model sits underneath all of it and never makes the headline slide, which is unfortunate, because it decides whether the person answering has seen your product before.
The distinction is simple to state. In a dedicated model, named agents are assigned to your account and stay there. In a shared model, agents work several accounts at once and pick up yours as calls arrive. Everything else follows from that one difference.
It is worth saying plainly that shared pools are not a scam. They are cheaper, they flex faster, and for straightforward high-volume work with short scripts they can be exactly right. The problem is that the two models get quoted against each other as if they were the same product, and the buyer usually cannot tell which one is on the table until the contract is signed and the first difficult call comes in.
The two models, and what actually differs
The Breakdown
Both models put a trained person on the phone. What separates them is whether that person accumulates anything about your business between one call and the next.
What you are buying
Dedicated: your own team
Your account only · Named agents · Consistent shifts · Onboarding before live contacts
You are buying continuity more than you are buying hours. The same people work your queue every day, so product knowledge, escalation judgment and familiarity with your regular customers keep building instead of resetting with each rotation. That takes time to pay off, which is the honest trade: a dedicated team is slower to stand up and costs more per hour at the start, and it gets better every month it runs. If your product is simple enough that month-two knowledge looks identical to day-one knowledge, you are paying for something you will not use.
What you are buying
Shared: a multi-client pool
Several accounts per agent · Rotating coverage · Script-led · Lower entry price
You are buying elasticity and a lower rate. Volume spikes get absorbed by a floor that is already staffed, you are not paying for idle capacity overnight, and you can be live in days rather than weeks. What you give up is depth. An agent covering a dozen brands can follow your script accurately and still miss that the caller is one of your largest accounts, or that the question they just asked is the one that always precedes a cancellation. For password resets and order status that gap costs nothing. For anything requiring judgment, your customers hear it immediately.
What changes in practice
Three things move when you switch models, and none of them appear in a coverage table.
Knowledge depth. A shared agent learns the top twenty questions. A dedicated agent eventually learns the twenty-first, which is the one that was going to escalate.
Continuity. Shared pools rotate, so the customer who called last week explains the situation again this week. Dedicated teams hold the thread, and repeat callers notice faster than any survey will tell you.
Escalation judgment. Knowing when to break the script is the part you cannot write into one. It comes from having seen the pattern before, which requires having been there before.
How each one is priced
The models are usually quoted on different units, which is what makes them so hard to compare on a spreadsheet.
Dedicated: you buy capacity
Priced per agent or per full-time equivalent. You are reserving people, so you pay for the seat whether or not it is busy at three in the afternoon. Predictable monthly cost, and the number moves when you add heads rather than when volume moves. The efficiency question worth asking yourself is what your occupancy will actually be, because you are funding the gaps too.
Shared: you buy consumption
Priced per interaction, per minute, or on a blended seat you do not own outright. Cheap at low or spiky volume because you are not paying for idle time. The cost is less predictable, and it climbs with success. Plenty of businesses have moved to dedicated not for quality reasons but because per-minute billing stopped being cheap once volume grew.
Why the headline rates mislead
An hourly rate only means something once you know what sits inside it. Before comparing two numbers, get both vendors to say in writing whether the price includes onboarding and product training, call recording, agent scoring, CSAT tracking, monthly reporting, and after-hours coverage. On shared-pool quotes several of those are commonly line items you negotiate separately, and a rate that looked twenty percent cheaper can close most of that gap once the add-ons are in.
The other thing worth pricing is the work that does not appear on either quote: the hours your own team spends managing the vendor, rewriting scripts, and re-explaining the product after a rotation.
How to tell which model you're being sold
Before You Sign
Proposals rarely use the word "shared". They say flexible, scalable, blended, or simply do not mention staffing at all. Six questions settle it, and the useful signal is often how quickly you get a straight number rather than the number itself.
How many other accounts will my agents work?
The whole thing in one question. Dedicated is zero. Anything phrased as "primarily yours" or "typically one to two" is a shared pool with better wording.
How long is onboarding before anyone takes a live contact?
Days means a script. Weeks means product training. Ask what happens in that time, who writes the material, and who signs off that an agent is ready.
Can I know who my agents are?
Not a vanity request. If the vendor can name the team, introduce them, and tell you their shifts, the assignment is real. If the answer is that staffing is managed dynamically, you have your answer.
Is quality scored across your whole floor, or against my standards?
Shared operations tend to sample quality across all accounts and report an average. That average can look healthy while your specific queue is the one dragging it down, and you would not be able to see it.
What happens to my queue at peak?
Ask where your calls sit when three clients spike at once. In a dedicated model your team is yours and the question is whether you flexed enough headcount. In a shared model you are competing for the same agents, and priority is a contractual matter rather than a technical one.
What is agent tenure and turnover on my account specifically?
Continuity is the thing you are actually paying extra for, so it is fair to ask for the number. A vendor who tracks it per account will tell you. A vendor who quotes a site-wide figure is telling you something too.
CX Readiness Checklist
The Checklist
Fifteen questions across the five things that tend to decide whether your customer experience is working. Tick the ones you can answer without going and asking someone, and the score updates as you go. Use it yourself or hand it to your team.
Tick what you can answer without checking. The blanks are the gaps.
Abandonment Rate
Do you know what percentage of callers hang up before reaching anyone?
If nobody can give you the number, it isn't being managed. It is also the metric most likely to be quietly excluded from a vendor report.
Do you know when in the queue they hang up?
Abandons in the first fifteen seconds are usually misdials. Abandons at two minutes are customers deciding you aren't worth the wait.
Does anyone call the abandoned numbers back?
Most don't. It is the cheapest recovery available and it almost never has an owner.
After-Hours Coverage
Do you know when your calls actually arrive?
Coverage built around your office hours rather than your customers' behaviour leaves a gap that turns up in reviews before it turns up in reporting.
Does after-hours do anything beyond take a message?
Voicemail at 9pm and a callback at 10am is business-hours support with extra steps.
Could you double coverage inside a month?
Launches, outages, and seasonal spikes don't wait for a recruiting pipeline.
First-Contact Resolution
Do you know your first-contact resolution rate?
The single number that tells you whether people are being helped or merely answered.
Is it broken out by issue type?
One blended figure hides the two or three categories doing most of the damage.
Can your agents resolve without asking permission?
Plenty of teams have the knowledge and not the authority. The queue looks identical either way.
CSAT Tracking
Are you measuring satisfaction at all?
If the answer involves the phrase "we'd hear about it if there was a problem", you are relying on the small minority who bother to complain.
Can you see CSAT by channel and by issue type?
An overall score of 4.3 can hide a chat queue at 3.1. Averages are where problems go to hide.
Does anything happen when a score comes back low?
A number nobody acts on is a report, not a measurement.
Visibility
Could you tell me how your customers were treated last week?
Not the volume. The treatment. Most businesses cannot answer this, which is why churn tends to arrive as a surprise.
Are calls recorded and actually reviewed?
Recording without scoring is storage. Somebody has to listen for it to count.
Are agents scored against your standards or a generic rubric?
A vendor's house quality form measures politeness. It does not measure whether your product was explained correctly.
Frequently asked questions
A dedicated customer service team is a group of named agents assigned to one client account and no others, working consistent shifts. Because the same people handle the same queue every day, product knowledge, escalation judgment and familiarity with repeat customers accumulate over time rather than resetting when staffing rotates. Dedicated teams take longer to stand up and cost more per hour than shared staffing, and the value shows up in the months after go-live rather than in the first week.
A shared agent pool is a staffing model in which agents support several client accounts at the same time, picking up whichever contact arrives next. It is the standard model across much of the customer service outsourcing industry. The advantages are real: a lower rate, faster launch, and elasticity during volume spikes because the floor is already staffed. The trade-off is depth, since an agent covering many brands has limited room to learn any one of them beyond a script.
The difference is whether your agents work only for you. Dedicated agents are assigned to your account alone, so knowledge and continuity compound. Shared agents split their time across multiple accounts, so each contact starts closer to zero context. Everything else people compare, including brand voice, escalation judgment, quality scoring and continuity for repeat callers, follows from that single structural difference rather than from the technology or the coverage hours.
Per hour, usually yes, because you are reserving capacity and paying for the seat whether or not it is busy. Total cost is less clear cut. Shared pools are typically billed per interaction or per minute, which is cheap at low volume and climbs as you grow, and several things bundled into dedicated pricing are often separate line items on shared quotes, including onboarding, call recording, agent scoring, CSAT tracking and reporting. Normalize both quotes to the same scope before comparing the rates.
Ask how many other accounts your agents will work. A dedicated answer is a number, and the number is zero. Answers such as primarily yours or typically one to two describe a shared pool in softer language. Follow up by asking how long onboarding runs before anyone takes a live contact, whether you can be told who your agents are and what shifts they work, whether quality is scored against your standards or sampled across the vendor's whole floor, and what happens to your queue when several clients spike at once.
Most contact centers plan against a target of under 5 percent, and figures in the 5 to 8 percent range are widely treated as acceptable depending on industry and call type. Treat those as planning rules of thumb rather than hard standards. More useful than the headline percentage is knowing when in the queue people hang up, since abandons in the first few seconds are usually misdials while abandons after a minute or two are customers deciding the wait is not worth it.
It tends to, because first-contact resolution depends on whether the person answering knows enough to finish the job and has the authority to do so. Agents who work one account accumulate the product knowledge that turns a transfer into a resolution. The gain is not automatic. A dedicated team without proper onboarding, current training as the product changes, and permission to resolve without escalating will produce much the same numbers as a shared pool.
Yes, and it is a common path. Businesses often begin with shared or overflow coverage to handle volume cheaply, then move to a dedicated team once per-interaction billing stops being economical or once the issues coming in start requiring real product knowledge. The switch costs you a fresh onboarding cycle, so it is worth asking at the outset what moving between models would involve rather than discovering the answer during a renewal.
Not sure which model you're on?
Get a free CX assessment. We'll walk the checklist with you and show you where the gaps are.
