Account based selling for small teams, without the enterprise stack
Account based selling was designed by teams with budget for it. What genuinely transfers to a team of four, what does not, and the rules that bind it.
Read
Nobody schedules a quarterly data cleanup because they enjoy it. They schedule it because six months of small omissions have made the pipeline number unusable, and the cleanup is cheaper than admitting that.
Every small sales team we have worked with has the same artefact somewhere: a spreadsheet export from the CRM, made by someone who no longer trusted the CRM. It is usually filtered to the twelve deals that person is personally sure about. It is usually more accurate than the system it came from.
That spreadsheet is the diagnosis. When the CRM stops being the place the answer lives, hygiene has already failed, and every downstream number inherits the failure. Forecasts, capacity planning, commission, the decision about whether to hire, all of it runs on records that a person has quietly stopped believing.
What follows is not a data governance programme. Six teams of three to eight people cannot run one and should not try. It is six rules, each small enough to hold on a bad week, which is the only kind of rule that survives.
Hygiene loses every argument it is framed as an aesthetic. It wins when it is framed as the three things it actually decides.
It decides whether the forecast is a forecast. A pipeline built on stages that people set by feel produces a number with no error bars, and worse, no way to find out where it went wrong afterwards. We have written separately about building a pipeline you can actually forecast, and every mechanism in it assumes the underlying records mean something.
It decides whether you can diagnose a loss. When a deal dies, the useful question is which stage it died in and what was true at the time. If stage changes were backfilled in a rush before a pipeline review, that history is fiction and the five places deals actually die stay invisible.
And it decides your exposure. This is the part small teams routinely miss, so it gets its own rule below.
Shared ownership is the most common single cause of rot. A record owned by "sales" is owned by nobody, and the follow-up that everyone assumed somebody else had done is the follow-up that never happened.
The rule is mechanical: every account and every open opportunity has exactly one named owner, and reassignment is an explicit act with a date on it. When someone leaves, their records get reassigned in one sitting rather than inherited by ambient assumption. This costs about twenty minutes per departure and saves the quarter after it.
Write down, in one line each, what has to have physically happened for a deal to be in each stage. Not "buyer is interested" but "buyer has named the problem in writing and we know who signs". The test is whether two people looking at the same deal would put it in the same stage without conferring.
If they would not, the stage definitions are opinions and the pipeline is a mood board. Our lead qualification framework is built the same way and for the same reason: an event you can point at survives an audit, and a judgement does not.
The practical consequence is that stage regression must be allowed and must be normal. A team that treats moving a deal backwards as an admission of failure has taught itself to lie forwards.
Contact data goes stale at a rate nobody estimates correctly, and the platforms that consume that data have already put numbers on it. Google's Customer Match documentation is unusually concrete here: it states that "Customer Match lists have a maximum membership duration of 540 days" and that "to stay eligible, a list must have at least 100 members added or updated within the last 540 days". You can read the eligibility rules in Google's own Customer Match documentation.
You do not have to adopt 540 days. The useful move is to adopt some number, apply it as a field on the record, and let it expire without discussion. A contact not verified inside your window is marked stale automatically. No meeting decides it, which is the entire point, because the meeting is where good data goes to be argued about and left alone.
This is the rule that reframes the whole exercise, and most small teams have never had it put to them plainly. If you hold personal data on prospects, keeping it accurate and not keeping it forever are legal obligations in most of the regimes you are likely to sell into.
Article 5 of the GDPR sets out the principles directly. Personal data must be "adequate, relevant and limited to what is necessary in relation to the purposes for which they are processed". It must be "accurate and, where necessary, kept up to date", and the text goes further: "every reasonable step must be taken to ensure that personal data that are inaccurate... are erased or rectified without delay". And it must be "kept in a form which permits identification of data subjects for no longer than is necessary for the purposes for which the personal data are processed". The full text of Article 5 and its processing principles is short and worth reading once in the original.
Read those three together and they describe a hygiene policy. Minimise the fields, keep them current, delete what has aged out. A team that does this because it is the law ends up with a cleaner CRM than a team that does it because a manager asked nicely, and the work is identical.
Whether the GDPR itself binds you depends on where your buyers are, and that is a question for counsel rather than for us. The principles are a sound default regardless, because the regimes that have followed it broadly rhyme with it.
Required fields multiply during good quarters and get abandoned during bad ones. The end state is thirty fields of which four are trustworthy, and nobody knows which four.
Cut to the fields that change a decision. For most small teams that is owner, stage, next step with a date, expected value, source, and the last verified date from rule three. If a field has never been used in a decision, it is documentation of an intention, not data. Delete it.
The same discipline applies to what you report on. Our note on the six numbers that earn a place on a B2B dashboard makes the argument from the reporting end, and it lands in the same place: a small number of fields you trust beats a complete schema you do not.
The quarterly cleanup is a symptom. It exists because nothing catches drift as it happens, so it accumulates until it is visible from management height.
Replace it with a fixed ten minutes at the end of the week, per person, on their own records: every open deal has a next step with a date, every stage still matches its definition, anything untouched for a month gets closed or explicitly parked. Ten minutes across five people is under an hour a week and it removes the quarterly day entirely.
Tooling helps less than people expect at this size. If you are considering building something to enforce it, the honest analysis of what live data actually costs is worth reading first: rivl's note on what "real-time" really means for a sales dashboard makes the case that most teams asking for live data need current data, and that the gap between the two is a large bill. The same logic applies to hygiene automation. The weekly ten minutes is free.
Two things are worth saying rather than implying.
Hygiene does not create pipeline. A perfectly maintained CRM with nothing in it is a very clean empty room. If the constraint is that not enough conversations are happening, none of the above is the priority and doing it thoroughly is a comfortable way to avoid the harder work.
And hygiene has a ceiling below which it is not worth formalising. If you are two people who speak every day, the shared context in your heads genuinely is more accurate than any system you would maintain, and the rules above are overhead. The point at which they start paying is roughly when a deal can be worked by someone who was not in the original conversation. Before that, write the stage definitions down and skip the rest.
We audit the records behind the forecast and report what the number would be if only the defensible deals counted. Thirty minutes to scope it.
Book a Free Consultation