Why your own costs beat any rate card
A rate cardA seller's published price list for advertising space, airtime or creator work, treated by buyers as an opening position, not the price paid. is a seller's opening position, and the negotiation and insertion orders guide treats it as a starting point rather than a price. A benchmark from elsewhere is an average of buys the business did not make, on audiences it may not share, in conditions it cannot see. Neither tells a buyer whether a quote is good for this audience in this season.
The business's own history can. Every row describes a buy it actually made: what it asked for, what it paid on full cost, what it got against its own target audience, and why the price was what it was. Read across seasons, the rows show which media have delivered the target at what cost, which lines under-delivered, and how much booking late or buying in a peak week moved the price. That is the evidence the building a media plan tutorial asks for when it says a plan should start from past performance rather than rate cards.
The history is not a budget tracker. The guide to tracking spend against budget and the tutorial on setting up a marketing budget tracker record money committed, spent and invoiced against budget lines. The cost history records what each unit of media cost and delivered, so it can be compared with the next quote. The two share source documents and should agree on totals.
It is not the post-campaign analysis either. The post-campaign analysis guide reviews one campaign against its plan; the cost history collects the cost and delivery rows from every analysis into one place so they can be read together.
Before you start
- The orders. Insertion orders, booking confirmations and platform campaign settings for every line bought in the period you will cover, with their pricing models and agreed rates.
- The invoices. Final invoices from media owners, agencies and technology providers, including fees.
- Delivery evidence. Reconciled delivery reports, airing and posting logs, ad server counts and any verification reports, as the measurement guides describe.
- The plans and post-campaign analyses, for the target audience each line was bought against and the delivery it achieved.
- A spreadsheet or database someone in the business owns, with a column structure agreed before the first row.
- The buying models and pricing guide, for the conversion to a common basis and the fees that sit outside a quoted rate.
The steps
Decide what one row records
One row is one line bought: one medium, one seller, one placement or package, one period. Do not record whole campaigns as single rows, because a campaign mixes media whose costs cannot be averaged meaningfully. Do not split a line into weeks unless the price changed between them. Write down the rule and the columns before entering data; the table after these steps lists the fields most histories need. A history whose rows mean different things in different years cannot be read.
Collect the orders, invoices and delivery reports
Gather the documents for each line and keep a reference to each in the row. The order says what was bought and at what agreed price; the invoice says what was charged; the delivery evidence says what ran. Where the three disagree, use the reconciled figures from the post-campaign analysis or the reconciliation the auditing guide describes, not the seller's first report. Start with the most recent campaigns and work back as far as the documents allow; a partial history that is accurate is worth more than a complete one built on guesses.
Record the pricing model and the full cost
Note the pricing model each line was bought on: per thousand impressions, per click, per action, per spot, per site per period, per insertion, or a fixed fee. Then record the full cost to the business: media cost, agency commission or fee attributable to the line, technology and data fees, verification and ad serving, and any surcharges. The buying models guide reports that the IAB has published work on supply chain transparency setting out where programmatic fees arise; record those fees in their own columns rather than folding them into the media price, so a later quote with different fees can be compared fairly. Note whether each price was gross or net.
Record delivery against the target audience
Record what each line delivered: ImpressionOne display of an ad or search listing, as counted by the platform. counts, spots aired, sites posted, insertions run, clicks or actions, as the pricing model requires. Then record the share of that delivery estimated to reach the target audience, with the source of the estimate. Where the line had a ReachUnique people who saw an ad. estimate, record it with its source. Delivery against the total audience flatters broad media; delivery against the target is what the next plan needs. If the target share is unknown, say so in the row rather than assuming it was all on target.
Convert every line to a common basis
Divide each line's full cost by its target impressions and multiply by a thousand to give cost per thousand target impressions (CPMCost per 1,000 impressions.), as the buying models guide sets out. For spots, sites and insertions, the target impressions come from the audience estimate behind the buy; the impression calculator and the cost per thousand calculator handle the arithmetic. Keep the native rate as well, such as CPCAverage cost of each click. or CPACost of each conversion. for performance lines, because native rates are what the next quote will arrive in. The common basis is for comparing media; the native rate is for comparing like with like.
Note the conditions that moved the price
Price depends on circumstances as much as on the medium. For each row, note what was true when it was bought: how far ahead it was booked, whether it ran in a peak or quiet period, whether it was guaranteed or non-guaranteed, its route (direct, through an agency, or a programmatic deal of the kind the programmatic deal types guide describes), any added value, any make-goodsReplacement delivery a media owner provides when a booked spot, insertion or placement fails to run as agreed, offered under the order's terms., and anything unusual such as a distressed late offer. These notes are what make a row comparable with a future quote. Two identical cost figures bought in different conditions mean different things.
Tag each row by medium, period and route
Add consistent tags so the history can be filtered: medium, format, seller type, audience definition, campaign, season, route and objective. Use a fixed list for each tag rather than free text, so a filter for one medium finds every row. Tag the audience definition carefully: a cost per thousand against one audience cannot be compared with one against another, and the tag is how a reader knows which rows belong together.
Read the history before planning and negotiating
Before a plan is built, filter the history for each candidate medium and audience and read the range of costs achieved, not a single average, along with the conditions behind the highest and lowest rows. Use the range as the plan's cost assumption and state that it came from the history. Before a negotiation, read the rows for that medium and seller type to know what a fair price looks like for this audience in this season. Work published through the IPA, as the negotiation guide reports it, stresses that good negotiation produces deals both sides can deliver; the history is for fair, evidenced challenges, not for squeezing a seller below what it can sustain. Where rows are few, read them as the guide to reading small numbers advises: as a range with wide uncertainty.
Keep it current after every campaign
Add the rows for each campaign as part of its wrap-up, once delivery and invoices are reconciled, and before the post-campaign analysis is filed. Give one named person the job. Each year, check that the tags and definitions still mean what they did and note any change in measurement source, so that a change in recorded cost is not mistaken for a change in price. A history that stops being updated becomes a rate card of its own, and an out-of-date one. At the end of each season, the same rows let the business place every seller's prices within its own range when reviewing its media owners.
The fields a cost history needs
The table lists the fields most histories need. Add columns for anything the business buys often that the table does not cover, and keep the definitions in a note at the top of the file.
| Field | What it holds | Why it matters |
|---|---|---|
| Line reference | Campaign, line name, order and invoice references | Traces every figure back to a document |
| Medium and format | Medium, format, length or size as bought | Groups comparable buys |
| Seller type and route | Media owner, agency or platform; direct, agency or programmatic deal type | Route changes cost and fees |
| Period and booking lead | Dates run and how far ahead the line was booked | Season and lead time move price |
| Audience | Buying audience, target audience, source of the size estimate | Cost is only comparable on the same audience |
| Pricing model and rate | Model, agreed rate, gross or net | Lets the next quote be compared in its own terms |
| Full cost | Media, commission or fee, technology and data, verification, surcharges, each separately | Shows what the line really cost |
| Delivery | Impressions, spots, sites, insertions, clicks or actions as relevant, reconciled | What the money bought |
| Target delivery | Target impressions and reach where known, with source | The basis for comparison |
| Common basis | Cost per thousand target impressions on full cost | Puts every line on one scale |
| Conditions | Guaranteed or not, added value, make-goods, peak or quiet, late or early | Explains why the price was what it was |
| Outcome measures | Native outcome rates for performance lines, against their own target | Adds result to cost for lines judged on outcomes |
The media planning framework lists the planning decisions the history feeds, and the media buying hub sets out the routes and pricing models the tags describe.
A boat hire firm builds its cost history (illustrative)
Ulmorvex Boat Hire builds three seasons of cost history
Ulmorvex Boat Hire, an invented river boat hire firm, has bought radio, regional press, paid social and paid search over three summer seasons, mostly booked by its marketing manager. For its next plan it builds a cost history from the orders, invoices and delivery reports on file: 46 rows across the three seasons.
Converted to cost per thousand target impressions on full cost, against adults aged 30 to 60 within an hour's drive, radio ranges from £7.80 to £11.40, regional press from £14.20 to £19.60 and paid social from £5.10 to £8.70, the social rows including platform fees and a verification charge recorded separately. Paid search is kept on its native measure: cost per booking ranges from £18 to £27.
The conditions column explains most of the spread. The two most expensive radio rows were booked less than two weeks before the peak bank holiday weeks; the cheapest was booked eight weeks ahead for early-season weeks. The most expensive press row was a premium position the firm has not repeated.
When a station quotes £13.50 per thousand target listeners for June, booked six weeks ahead, the manager shows the station that the firm's own rows for similar conditions sit between £8.20 and £9.60, and asks what in this quote is different. The station revises the daypart mix and the quote falls to £10.10. For the plan itself, the manager sets each medium's cost assumption at the middle of its range for comparable conditions and notes the source as the firm's cost history.
For lines bought by the spot, site or page, the walk-through of the CPM calculator for offline media derives the full cost and target exposures each row needs before conversion.
Common mistakes
- Recording campaigns instead of lines. A campaign average mixes media that cannot be compared. One row per line bought.
- Recording the rate, not the full cost. Fees, commissions and verification change the comparison; record them separately and include them in the common basis.
- Using total audience. Cost per thousand against the total audience flatters broad media. Record target delivery with its source.
- Leaving out the conditions. A cost figure without its booking lead, season and route cannot be compared with a new quote.
- Treating one row as the norm. A single unusual buy, such as a distressed late offer, is not a price. Read ranges.
- Using unreconciled delivery. First reports change once make-goods and credits are agreed; record the reconciled figures.
- Letting it lapse. A history updated once and then abandoned becomes a stale rate card. Add rows after every campaign and give the job to a named person.
