What a pilot has to prove
A pilot is not a small version of the full programme run in the hope that it works. It is a test of three claims, and it should be designed so that each can be answered. First, that sellers and marketing can run coordinated plays on named accounts with the time they actually have. Second, that the plays move accounts: more of each buying committeeThe group of people in an organisation who influence, evaluate and approve a purchase. reached, more engagement from the right roles, more meetings. Third, that the movement turns into pipeline and better deals than the same kind of account produces without the plays.
The third claim is the one most pilots cannot answer, because they have nothing to compare against. Accounts chosen for fit tend to progress anyway; a pilot that reports only its own accounts takes credit for deals that would have happened regardless. The business-to-business measurement guide names holding out account groups as one of the few methods that show whether marketing caused anything, and the incrementality testing guide explains the logic of a holdout groupA randomly chosen part of an audience kept from exposure to measure what exposure caused.: comparable before the test, untouched by the treatment during it, and read against the treated group afterwards. A pilot's comparison accounts are that group.
| Measure | What it shows | When to read it | Compared with |
|---|---|---|---|
| Committee coverage | Roles reached in each account, counted as roles rather than contacts | Monthly | The same count in comparison accounts |
| Engagement by role | Whether the right people are responding to the plays | Monthly | Comparison accounts, and the pilot's own first month |
| Meetings held | Whether engagement is turning into conversations with sales | Monthly | Comparison accounts |
| Opportunities and pipeline | Whether conversations become deals | At the end of the cycle | Comparison accounts |
| Win rateShare of closed sales opportunities that were won. and sales cycle lengthAverage time from opportunity creation to a closed deal. | Whether the deals are better, not only more numerous | As deals close, often after the pilot | Comparison accounts and the segment's history |
Read every figure in that table as the guide to reading small numbers advises. A short list gives a direction, not proof: show the counts, put ranges around the shares, and make no claim of significance from a few dozen accounts. The decision at the end is a judgement informed by the comparison, which is still far better than a judgement informed by nothing.
Before you start
- The account-based marketing guide read in full, since this tutorial uses its tiers, committee mapping and plays without restating them. The comparison of ABM and demand generation helps confirm that the market suits a named-account approach at all.
- An ideal customer profile the sellers have signed off. The tutorial on building an ideal customer profile from CRM data produces one; the pilot draws its segment from it.
- Named sellers who have agreed to take part, with their managers' agreement that pilot accounts are part of their job rather than extra to it.
- A CRM that records accounts as well as contacts, with a field to mark pilot and comparison accounts and engagement visible at account level.
- A sales cycle length for the segment, read from the business's own closed deals, because it sets how long the pilot runs before its result can be judged.
- A written decision rule, agreed in advance: what the comparison would need to show for the team to scale, to extend, or to stop.
The steps
Choose one segment and the sellers who will work it
Pick a single segment from the ideal customer profileA description of the type of organisation most likely to buy, succeed with the product and stay, used to choose target accounts.: one industry, one size band, one problem the product solves well. A pilot that spans several segments needs different plays for each and ends with too few accounts in any of them to read. Then name the sellers who will work the accounts and ask each how many accounts they can work at tier-two depth alongside their other duties, as the account-based marketing guide's advice on sizing the list from what named people can work suggests. Their answer, added up, is the size of the pilot group. Record it, because it is the constraint every later step respects.
Build the fitting list with sales
With the sellers, build a list of accounts in the segment that fit the profile and are not already in an open opportunity or a customer. Make it long enough to supply both the pilot group and a comparison group of similar size. Sellers add the accounts they know fit and remove the ones they know do not, with a reason written for every removal. The academy exercise on building a target account list for an industrial sensors maker practises the same reading of fit. Accounts with existing relationships or recent engagement should be noted, because the split in the next step has to balance them.
Split it into pilot and comparison accounts
Divide the list into two groups that look alike before the pilot starts. A simple method is to pair accounts that are similar in size, sub-sector, past engagement and relationship with sales, then assign one of each pair to the pilot and the other to the comparison group by a coin toss rather than by preference. Letting sellers choose their favourite accounts for the pilot makes the comparison meaningless, because the favourites were always more likely to buy. Mark both groups in the CRM. The comparison accounts receive only the usual programme: the broad demand activity, inbound handling and ordinary sales prospecting they would have had anyway. They are not excluded from the business, only from the pilot's plays. If an agency runs part of the pilot, the guide to hiring and evaluating a B2B demand agency keeps the list, the audiences and the plays in the business's own hands, so the comparison survives a change of provider.
Map the committee for each pilot account
For every pilot account, map who decides, who influences, who uses, who blocks and who signs, with names where they can be found and gaps marked where they cannot. The account-based marketing guide sets out the roles and what each cares about, and the ABM account plan template holds the map, the plays and the sales actions for one account in one place. Keep the plans light: at tier-two depth the map matters more than a long narrative. Note which roles are already engaged, because committee coverage is the first leading indicator the pilot will read.
Choose two or three plays and write their rules
Choose a small number of plays at tier-two depth: personalised by segment and role rather than bespoke to each account. For each play, write the entry condition (what puts an account into it), the sequence of marketing touches and sales actions with who does each, and the exit condition (a meeting, an opportunity, or a decision to pause). Two or three written plays that sellers understand are worth more than a menu of ideas nobody runs. Keep every play's content and timing the same across pilot accounts, so that the pilot tests the plays rather than a different improvisation for each account.
Set up advertising to the pilot list
Advertising usually carries part of each play. As the LinkedIn account-based marketing guide explains from LinkedIn's Marketing Solutions Help, listed in this page's sources, an uploaded company list takes time to process and has to clear the documented minimum audience size after matching. A short pilot list on its own may not clear it. The usual answers are to run one campaign across all pilot accounts rather than one per account, to widen the audience with a role filter across a larger list, or to lean on channels that do not depend on a matched audience, such as email, direct outreach and events. If the list is widened, keep the comparison accounts out of it, or the comparison group stops receiving only the usual programme. The LinkedIn guide covers the build itself.
Agree what engaged means and how fast sales act
Write down what makes a pilot account engaged enough for a seller to act: for example, a set number of roles active within a period, or a specific action by a senior role. Then agree how quickly the seller acts when an account crosses that line and what the action is. The sales and marketing alignment guide explains why a service level only works when it is measured from CRM timestamps and reported in the same meeting both teams attend. In a pilot the service level is also part of the test: if sellers cannot act on engaged accounts in time, the pilot has found a capacity problem, which is worth knowing before scaling.
Run a weekly joint review
Meet weekly, briefly, with the pilot's marketer and sellers. Go through the pilot accounts that moved: which entered or left a play, which crossed the engaged line, what the seller did and what happened. Agree the next action for each and record it in the CRM. Note anything that changes the comparison, such as a comparison account that arrives through inbound and is worked normally, which is fine, or a seller who starts running a play on a comparison account, which is not. The weekly review is where the plays are kept running as written and where problems with content, timing or capacity show first.
Read the leading indicators each month
Once a month, read committee coverage, engagement by role and meetings held for the pilot accounts and the comparison accounts side by side. The point is direction: are the pilot accounts pulling ahead on the measures that usually come before an opportunity? If after several months they are not, look at the plays and the sellers' follow-up before concluding that the model does not work, and record what was changed and when. Avoid reacting to one month; small groups move around from month to month, and the experimentation loop is a better rhythm than a weekly verdict.
Compare the groups and decide to stop, extend or scale
At the agreed end date, after at least one sales cycle, compare the two groups on meetings, opportunities, pipeline value, and, as deals close, win rate and cycle length. Lay out the counts for each group, the cost of the pilot, and what the sellers spent their time on. Then apply the decision rule written before the pilot began. Scale when the pilot accounts are clearly ahead on pipeline and the sellers could run the plays at the pace required; extend when the leading indicators are ahead but the opportunity difference is too small to call; stop when neither group differs or the plays could not be run. Write the decision down with the evidence. If the programme is scaled and later stalls, the diagnosis of account-based marketing not working is the place to start.
A pilot's job is to earn the larger programme or to save the business from building it. The demand section sets account-based work beside the broad demand programme that most teams keep running alongside it.
A lab software vendor's first pilot (illustrative)
Ottercrane Lab Systems pilots plays on contract research labs
Ottercrane Lab Systems is an invented vendor of laboratory information software. Its profile names contract research labs with several sites, and its sales cycle in that segment runs about nine months. Two sellers say they can each work 15 accounts at tier-two depth, so the pilot group is 30.
Marketing and the sellers build a list of 64 fitting labs, remove 4 already in open opportunities, pair the remaining 60 by size, sites and past engagement, and assign one of each pair to the pilot by coin toss. The 30 comparison labs keep receiving the newsletter, the quarterly webinar and normal inbound handling. The pilot runs three plays: a sector report sent to quality and lab operations leads, LinkedIn advertising run as one campaign across all 30 pilot accounts with a separate message for each role, and a peer-customer webinar followed by seller outreach to any account with three or more roles engaged. The sellers agree to act on an engaged account within two working days.
| After nine months | Pilot accounts | Comparison accounts |
|---|---|---|
| Accounts with three or more committee roles engaged | 19 of 30 | 7 of 30 |
| Accounts with at least one meeting held | 16 | 8 |
| Accounts with an opportunity opened | 6 | 4 |
| Opportunities won so far | 2 | 1 |
| Programme cost beyond the usual programme | £26,000 | none |
The leading indicators are clearly ahead. The opportunity difference, 6 against 4, is too small to call either way, and most of the opportunities have not closed. Under the rule written at the start, that result means extend, not scale: the team runs a second cycle with the same split and the same plays, adds a quarterly check on whether the comparison labs have been contacted outside the usual programme, and agrees to read win rate and cycle length once the open opportunities close.
Common mistakes
- Sellers choosing the pilot accounts. The favourites were always more likely to buy, and the comparison measures the choice rather than the plays.
- No comparison group. Without one, the pilot credits itself with the progress that fitting accounts make anyway.
- A list sized by ambition. A pilot the named sellers cannot work tests capacity rather than the model.
- Several segments at once. Each needs its own plays, and each ends with too few accounts to read.
- Plays improvised per account. The pilot then tests a dozen different things once each, and nothing can be repeated if it works.
- Advertising to a list too small to run. Check the platform's documented minimum after matching before the plan depends on it, and keep comparison accounts out of any widened audience.
- Judging after one quarter. Leading indicators move first; opportunities need at least a cycle, and closed deals often arrive after the pilot ends.
- Deciding the rule afterwards. A stop, extend or scale rule written once the numbers are in will always find a reason to scale.
