Why a vague brief costs more than a precise one
The most expensive mistake a call centre manager can make when buying data is submitting a brief that says "SME decision-makers, UK, 10,000 records." That description fits three or four entirely different products at three or four different price points, and a supplier who cannot narrow it further will default to the most generously-interpreted file they hold. You will pay for data your dialler cannot use.
A complete brief specifies six parameters. Get all six right and you will receive a file that loads cleanly into your dialler, clears compliance, and does not waste 40% of records on numbers that never connect. Miss one and you are chasing the supplier for replacements by week two.
The six parameters are: audience definition, required fields, volume cadence, TPS or CTPS wash recency, data freshness window, and licence terms. The rest of this article works through each in detail, then covers pricing, replacement guarantees, and the QA checks to run before a campaign goes live.
The six brief parameters explained
1. Audience definition: SIC code for B2B, demographic for B2C
For B2B campaigns, target by UK Standard Industrial Classification 2007 (SIC 2007) code, not by vague sector labels. "Manufacturing" covers 70+ SIC codes; a Leeds-based precision engineering firm targeting Operations Directors in food manufacturing should specify SIC 07, 10, and 11 rather than leaving the supplier to interpret "food sector." Add employee band, turnover band, and geography (postcode district or county) to reduce the count to something workable.
For B2C campaigns, define the demographic profile precisely: age range, household income band, property tenure (owner-occupier vs renter), presence of children, or a specific interest flag such as home improvement intent or travel frequency. Consumer files built on consented opt-in data carry interest and lifestyle flags at the record level; without those flags in the brief, the supplier has no instruction for how to filter.
Named contact versus role-based is a separate choice. A named contact ("James Hargreaves, Head of Procurement, Hargreaves Logistics, Bradford") carries more credibility on the call but decays faster. Role-based contacts ("Head of Procurement") remain dial-able longer but require agents to work through some gatekeeper friction. Most call centres handling complex B2B sales take named contacts; outbound lead generation at scale often accepts role-based to maximise count.
2. Required fields: direct dial, mobile, or switchboard
This single decision shapes your cost per record more than any other. Direct-dial B2B numbers route straight to the decision-maker and typically cost £0.20 to £0.50 per record. Switchboard numbers cost £0.05 to £0.15 but require agents to get through to the right person, adding 60 to 90 seconds per attempt and cutting effective contacts-per-hour by 25 to 35%.
Mobile numbers for B2B contacts are the premium tier. They carry the highest connection rates (typically 30 to 50% higher than landline) and are listed as direct contact on the record. For roles that are rarely desk-based (field sales managers, logistics coordinators, site managers), mobile-only targeting is often the only viable option. See our detailed guide on B2B direct-dial numbers in the UK for connection-rate benchmarks by field.
For B2C, the landline versus mobile split matters for both compliance and connection rate. Landlines have higher TPS registration rates among older demographics; mobile-only consumer data suits younger audiences and avoids that suppression loss. Specify both if your dialler supports mixed-mode dialling.
3. Volume cadence: matching data to dialler throughput
Ordering 100,000 records and loading them all on day one is the sign of a brief that never asked about dialler capacity. The right volume question is: how many dials per agent per day, multiplied by how many agents, over how many weeks?
A predictive dialler running at 4:1 ratio (four dial attempts per contact) with 20 agents handling 300 dials per agent per day consumes 6,000 dial attempts daily. At a 20% connection rate, that produces around 1,200 live conversations. If the average contact is called twice before being retired or converted, you need 600 fresh records per day just to keep pace. Order 12,000 records per fortnight, not 100,000 upfront.
Phased delivery also protects quality. Receive the first tranche, run a QA pass (see below), confirm connection rates match the guarantee, then release subsequent tranches. Suppliers who push back on phased delivery should raise a flag.
4. TPS and CTPS wash recency: the 28-day rule
The Telephone Preference Service (TPS) covers personal UK telephone numbers; the Corporate Telephone Preference Service (CTPS) covers numbers registered by businesses and organisations. Under the Privacy and Electronic Communications Regulations (PECR), calling a TPS- or CTPS-registered number without the individual's prior consent is a civil contravention that the Information Commissioner's Office (ICO) can prosecute, with fines up to £500,000 for serious cases.
The ICO and industry bodies treat 28 days as the maximum acceptable gap between a TPS wash and the date of dialling. Critically, that 28-day clock runs from the point of the call, not the point of data delivery. If a supplier washes their file monthly and you receive the data on day 27 of their wash cycle, you have one day before the suppression is stale. Ask specifically: "What was the TPS wash date on this file, and what process do you have for re-washing before delivery?"
Suppliers who wash their master file on a rolling 28-day schedule and certify the wash date at point of delivery are the standard you should insist on. For more on TPS and CTPS compliance in practice, read our article on UK telemarketing TPS and CTPS rules.
5. Data freshness window
B2B data decays at roughly 25 to 30% per year. An employee band changes, a decision-maker is promoted or leaves, a number is reassigned. For a call centre, stale data translates directly to wasted dials. The working standard is to specify records verified or updated within the last 12 months for B2B and within 18 months for B2C.
Freshness is not the same as the supplier's file compilation date. A supplier may hold a large master file and only re-verify records on a rolling schedule. Ask: "For records in my SIC code and geography, what proportion have been telephone-verified within the last 12 months?" A supplier who cannot answer that question numerically probably cannot enforce the freshness window you need.
Some call centres buying for mid-market or enterprise B2B campaigns specify records verified within 6 months. That cuts volume significantly but raises connection rates. In our experience, for campaigns targeting C-suite contacts at firms with 50 or more employees, the higher-freshness specification is worth the reduction in count: you are paying per meaningful conversation, not per record.
6. Licence terms: single-use, multi-touch, or perpetual
A single-use licence permits one outbound campaign. You dial the file once and retire it. This is appropriate for a one-off promotion, a product launch, or a seasonal campaign where you will not need to re-contact the same prospects.
A multi-touch licence allows multiple contacts over a defined period, typically three to twelve months. This suits call centres running follow-up sequences (initial call, callback, qualification call) or combined call-plus-postal cadences. Multi-touch pricing is typically 1.5x to 2.5x single-use for the same file.
Perpetual licences transfer the right to hold and dial the data indefinitely. They are rare for telemarketing data specifically because the data decays and TPS compliance must be re-verified continuously. Perpetual licences are priced substantially higher and are generally more appropriate for CRM enrichment than for active dialling files. For most call centres, a 12-month multi-touch licence is the practical maximum before a re-purchase and re-wash is the correct operational move.
Predictive versus preview dialler: how the choice changes your brief
Predictive diallers generate outbound calls automatically and connect agents only when a human answers. They are built for volume. A typical predictive setup processes 300 to 600 dials per agent per day and suits B2C lead generation, utilities upsell, and similar high-volume outbound where the marginal cost of a wasted call is low.
For predictive dialling, brief for breadth. You need a high record count, and connection rate matters more than named-contact precision. Switchboard numbers are acceptable because the agent time saved by automation absorbs some of the gatekeeper cost. Volume discount pricing (£0.05 to £0.12 per record) makes sense here.
Preview diallers show the agent the contact record before the call is placed. Agents review the record, decide whether to dial, and initiate the call manually. They are standard for regulated sectors (financial services, insurance, energy switching), complex B2B sales cycles, and any campaign where the call quality matters more than the call volume. Ofcom's persistent misuse rules on abandoned calls are worth reviewing for any predictive operation.
For preview dialling, brief for precision. Named contacts, direct dials, verified mobile numbers, and tight SIC code targeting. Pay more per record: £0.25 to £0.50 is normal. Order fewer records per tranche and expect higher conversion rates to justify the unit cost. The economics only work if each call has a realistic chance of progressing.
Pricing models and what drives the cost per record
UK call centre data is priced per record, not per campaign. The table below shows the main variables and their typical effect on price.
| Parameter | Lower cost end | Higher cost end | Typical range |
|---|---|---|---|
| Audience type | B2C volume file | B2B named contact, C-suite | £0.05 to £0.50 per record |
| Number type | Switchboard / landline | Direct dial / verified mobile | +40 to 80% premium for direct dial |
| Freshness window | Verified within 18 months | Verified within 6 months | +20 to 40% for tighter window |
| Licence term | Single-use | Perpetual | Multi-touch: 1.5x to 2.5x single-use |
| Volume ordered | 50,000+ records | Under 2,000 records | Volume discounts from 5,000 records upward |
| Geography | National file | Specific postcodes or counties | Narrower geo rarely adds cost; very small areas may have minimum order charges |
TPS suppression does not normally add cost: it is applied at fulfilment and the count you receive is post-wash. What changes is the count itself. A B2C consumer file can lose 15 to 25% of records after TPS suppression, so factor that into your volume brief. If you need 10,000 dialable numbers, order 12,000 to 13,000 raw and specify post-TPS delivery.
Replacement guarantees and QA before campaign launch
What a reasonable replacement guarantee looks like
Reputable suppliers back their data with a replacement or credit guarantee. The standard for B2B files is 10 to 15% replacement on records that are factually wrong at delivery: disconnected numbers, confirmed-gone employees at the named organisation, or duplicate records within the file itself. Claims are usually accepted within 30 days of delivery, backed by evidence from the dialler (call log showing the number is not in service, or a contact confirming the named person has left).
Ask two questions before signing: first, do replacements come from the same SIC code, geography, and seniority band as the originals? A replacement credit applied to a different sector is not a replacement. Second, does the guarantee cover records that TPS status changes after delivery? The answer should be no, because TPS is a live register and TPS additions after the wash date are the caller's responsibility to manage, not the supplier's.
QA checks to run before going live
Load the file into a staging environment first, not directly into the live dialler queue. Run these checks before any call is made:
- Duplicate check across telephone numbers (within the file and against your existing suppression list or CRM).
- Format validation: all numbers should be in E.164 format or your dialler's required format. Non-geographic numbers (0800, 0845) frequently cause issues with predictive diallers.
- TPS cross-reference: if your own TPS access is more recent than the supplier's wash date, re-wash before loading.
- Spot-check 50 to 100 records manually: does the company name match the postcode district? Does the named contact appear plausible for the stated role and SIC code?
- Count completeness check: confirm the delivered record count matches the invoice count, and that all specified fields (direct dial, mobile, email if ordered) are populated at the contracted fill rate.
A B2B file that passes these checks before launch will waste fewer dials and give you a cleaner baseline for measuring campaign performance. Trying to diagnose a poor connection rate mid-campaign without a pre-launch QA pass is far harder and wastes agent time.
Compliance obligations for the call centre, not just the data supplier
Buying data from a supplier who has compiled it lawfully does not transfer compliance responsibility to them. For B2B telemarketing, the calling organisation is the data controller under UK GDPR Article 6(1)(f) and must complete its own Legitimate Interests Assessment (LIA) before the campaign starts. The LIA documents why the call centre (or its client) has a legitimate interest in contacting this audience, why that interest outweighs the data subject's right not to be contacted, and what safeguards are in place (clear identification at the start of the call, a simple opt-out mechanism, suppression list maintenance).
For B2C campaigns, PECR consent requirements are stricter: unless the individual has previously given consent to be contacted by the specific organisation, the TPS and CTPS suppression is mandatory and there is no legitimate-interests bypass for live calls to personal numbers.
Maintain a suppression file. Every "do not call" request from a dialled record must be honoured across all future campaigns, regardless of whether the data is re-purchased from the same or a different supplier. Failure to suppress known objectors is one of the more common routes to an ICO investigation. For a deeper treatment of quality standards on the telemarketing side, see telemarketing data quality in the UK.