Skip to main content
A reference set for a B2B or sales-led business, where marketing generates and qualifies leads that a sales team closes. Adopt what applies, rename anything your team already calls something else, and drop unused rows. Metric names use the lowercase, underscore-separated format the flat file expects.
This is a starting point, not a mandate. Definitions here follow common industry usage, cited below, but your business gets the final call on inclusions and exclusions, recorded in your data dictionary. In particular, agree the MQL and SQL thresholds with sales in writing; a lead qualification standard marketing sets alone doesn’t hold up past the first disagreement.

​
Funnel volume

Store the count at each stage. Every conversion rate between stages is derived, never stored separately. MQL and SQL, defined. A marketing qualified lead is a lead a scoring model, built from behavior, demographic, or firmographic signals, has flagged as ready for sales attention. A sales qualified lead has been vetted by an actual salesperson and judged ready for a direct sales conversation. The distinction matters because an MQL is a marketing judgment and an SQL is a sales judgment; conflating them is a common source of friction between the two teams. As of 2026, most B2B organizations treat the MQL as an early milestone worth tracking rather than the primary measure of marketing’s success, since an MQL alone doesn’t guarantee a sale.

​
Funnel conversion rates

A commonly cited MQL-to-SQL conversion rate sits around 13% at the median, with top-quartile programs closer to 28%, though this varies widely by industry and by how strict the MQL threshold is set. A lower threshold produces more MQLs and a lower MQL-to-SQL rate; a stricter one produces the reverse. Neither is inherently right, but the two teams need to agree on where the line sits.

​
Cost and efficiency

Cost per lead varies enormously by industry; a commonly cited 2026 median across B2B sits in the low hundreds of dollars, with top-quartile programs well below that and specialized or regulated verticals well above it. Use a citable industry figure only as a rough sanity check on your own number, not as a target, since your actual cost depends heavily on channel mix, competition, and how strict your lead definition is. Match spend to the period that generated the result, not the period the deal closed in. A deal closing this month was very often generated by leads, and spend, from months earlier. Dividing this month’s spend by this month’s closed deals credits current spend with previous months’ work; report cost_per_closed_won against the cohort that generated the underlying leads instead.

​
What this pack needs from the rest of the system

A metric list on its own does not produce a sales-led report. Three other things have to be set up the same way, and each is where a sales-led business diverges from the ecommerce default this framework otherwise assumes.
Set all four up before the first report, not after. Each one is cheap to add on the week you start and impossible to backfill: a deal that closed without its source fields stored cannot be attributed later, because the click that produced it is long gone.

​
Pipeline

pipeline_value and sales_cycle_days are marked as stored inputs even though both are calculated, which looks like it breaks the flat file’s rule against storing derived numbers. It does not: both aggregate deal-level rows the flat file never holds, so neither can be recalculated from anything in the file. Record in the data dictionary which system computed them and over what set. See the exception.
Pipeline velocity is the single number that best captures whether the pipeline is speeding up, slowing down, growing, or shrinking, since it combines volume, quality (win rate), size, and speed into one figure. A rising velocity from more opportunities and a rising velocity from a shorter sales cycle mean very different things operationally, so read it alongside its four inputs rather than on its own.

​
Sources

Definitions and formulas above follow common usage from the following, current as of 2026: Benchmark figures (cost per lead, conversion rates) are directional and vary widely by industry and lead-quality threshold; treat them as a reference point for a first target, not a fixed standard, and replace them with your own once you have a few months of real data.
Last modified on August 11, 2026