A day of CRM work

Definitions do not survive an interview. Here is an ordinary Tuesday, in the order it happens.

08:40 — the hygiene sweep, before anything else. You open a saved view called My open deals — no next step: opportunities where the field is empty or its date has passed. Two appear. This takes four minutes and is the highest-leverage four minutes of the day, because every deal on that list is drifting. Then No activity in 14 days: three deals. One you consciously parked, so you write that into the next step — "holding until their budget cycle opens in January — re-engage 6 Jan" — and it stops reading as neglect and starts reading as a decision.

Why "no next step" is the most damning field. Every other field can be defended. An optimistic close date is a judgement call; a stage might be arguable; an amount might be a placeholder. An empty next step is not a judgement at all — it is proof that your last conversation ended without agreeing what happens next. Either the buyer would not commit (information you are ignoring) or you did not ask (a skill gap). It is also not a next step if it is undated, or if only you do it: "send follow-up email" is a task; "results review with lab director and deputy CFO, booked 14 Oct" is a next step, because the buyer has spent diary time to be in it.

09:30 — the call. Thirty minutes with the lab director on the proof-of-concept results.

10:05 — logging it (the five-minute rule). Now, not Friday. Reps who batch CRM work to a Friday purge never catch up, and by then the detail has decayed to "good call, they're keen". A good logged call has four parts: who was on it and in what role; what they said in their own words ("turnaround time on histology is the metric the trust is measured on and we're missing it three days a week" beats "discussed pain points" — buyer language is what you reuse in the proposal); what changed about the deal (new stakeholder, constraint, budget, timeline, competitor); and what was agreed, with a date.

What does not go in: private assessments of individuals, speculation about people's jobs, anything about a patient, and anything you would not want the customer to read. CRM records are exportable by administrators, discoverable in disputes, and potentially disclosable to the person they describe if they make a subject access request.

10:15 — the stage, against exit criteria. The POC passed all five pre-agreed criteria and the lab director confirmed it in writing. Your company's exit criterion for leaving Evaluation is evaluation completed against success criteria agreed in advance, with the buyer confirming they were met. That is now true, so the deal moves to Proposal. Had you only a warm verbal from your champion, it stays put. Mature orgs enforce this with a validation rule; immature ones do not, which is exactly why the habit has to be yours rather than the software's.

10:20 — the close date. The trust's capital approval meeting is 12 November and paperwork must land ten days before, so the close date becomes 30 November — derived from their process. Not 31 December because it is the quarter end, and not last quarter's date left in place. A close date the buyer has never seen is the root cause of almost every serially slipping deal.

10:25 — the buying group. The lab director mentions that information-governance sign-off goes through a named IG lead you have never spoken to. You create them as a contact, attach them to the opportunity as a contact role (influencer, realistically a blocker), and add a task: identify route to IG lead via lab director, by Friday. This is not admin — it is the CRM expression of multi-threading, and it is one of the few sales behaviours with hard published evidence behind it.

Voice from the field

"A won $50K–$250K deal these days typically involves at least 10 stakeholders."
— Dan Morgese, Director of Content Strategy and Research, Gong (Gong Labs research on multi-threading)

Gong's analysis of the same data also found win rates dip when an evaluation opens with an executive but rise when executives join a little later — around the third touchpoint — a useful corrective to demanding the CFO on call one. Either way: if your opportunity carries one contact role, the record is telling you something true and uncomfortable.

10:30 — close-out. Amount unchanged. Forecast category stays best case, not commit, because IG has not started. Next step updated, task created. Total CRM time for a 30-minute call: about eight minutes.

16:00 — the pipeline review. Your manager opens a dashboard, not your notes: stage distribution, deals by close date, hygiene exceptions, week-on-week movement. When they land on your deal you talk in the Module 2 grammar — stage, value, close date, next step, risk — and every clause is already in the record they are looking at. That congruence is, in practice, what "trustworthy rep" means.

Field note

Every CRM you inherit will be dirtier than anyone admits: duplicate accounts, close dates set to 31 December of a year that has ended, next-step fields reading "follow up", opportunities owned by people who left. Nobody expects a new joiner to fix that, and offering to in week one reads as naive. What is noticed, quickly and by more people than you expect, is whether your records are clean. Within a quarter that converts into practical advantage: your deals get believed, your requests to ops get built, and your commission queries get settled from the record rather than from memory.

4. The wider stack

The CRM is the system of record. Everything else exists because the CRM is bad at something specific, and each tool earns its budget by feeding data back. Think of four rings: data in (who to talk to), engagement out (reaching them at volume without becoming a spambot), intelligence back (what actually happened), and paperwork through (turning agreement into a signed contract).

Sales engagement platforms (sequencers) — Outreach, Salesloft (which bought the conversational-marketing tool Drift in 2024, adding inbound website chat to its outbound engine), plus Apollo, HubSpot Sequences and Gong Engage. What it does: runs multi-step, multi-channel sequences or cadences — email day 1, call day 2, LinkedIn day 4 — queues the day's tasks, dials, tracks replies, pauses automatically when someone responds, and reports conversion by step and template. Why companies buy it: consistency and measurement. Without it, follow-up collapses after touch two, which is precisely where most replies come from, and nobody knows which message works. Daily rep use: work the task list, personalise the top-of-sequence emails, skip contacts research says do not fit, log call dispositions, and pull people out of automation the moment a human conversation starts.

Prospecting data. LinkedIn Sales Navigator is the professional-network layer: advanced lead and account search, saved lists, alerts on job changes and company news, InMail, and AI research summaries (Account IQ, Lead IQ). Because people maintain their own profiles, title data is fresher here than anywhere — but by design it does not hand you email addresses. ZoomInfo is the large contact-and-company database, strongest in the US; it leaned so far into go-to-market positioning that it changed its Nasdaq ticker to GTM in 2025. Cognism is the one to know for UK and European selling: its differentiator is compliance and phone-verified mobile data, screened against do-not-call registers in many countries including the UK's TPS and CTPS, with a notification programme for individuals whose data it holds — exactly what a UK data protection officer will ask about. Apollo combines database and sequencer with a usable free tier. Why companies buy it: rep time — a week of list-building becomes an afternoon — plus enrichment, automatically filling firmographic fields so segmentation and territory design work at all. Daily rep use: build and refine target lists against the ICP, watch alerts for trigger events, pull contact details, push good ones into the CRM and a sequence. Treat all of it as claims, not facts: bought data decays at roughly 2–3% a month, and opening a call with a title the person left eighteen months ago burns the call.

Conversation intelligence — Gong, Chorus (ZoomInfo), Clari Copilot, plus native tools inside Outreach and Salesloft. What it does: records, transcribes and analyses calls — summary, talk-to-listen ratio, topics, competitor mentions, whether a next step was agreed — and rolls it into deal-risk signals and a searchable library. Why companies buy it: coaching at scale (review ten calls in the time it took to sit in on one), deal risk (flagging that no decision-maker has appeared in six weeks), and institutional learning (marketing and product hear real objections rather than a rep's paraphrase). Daily rep use: push the auto-summary into the CRM instead of retyping; clip a two-minute segment for a colleague; search how the best rep handled the objection you just met; before a call with a new account, listen to the last one anyone had with them. For a career changer this category is a gift — your first fortnight can include fifty real calls, the fastest available substitute for experience. One etiquette point: recording requires transparency, not stealth. Announce it, offer the option to decline, and accept refusal instantly — some organisations, NHS bodies among them, will refuse outright.

Scheduling — Calendly, HubSpot Meetings, Chili Piper. Publishes real availability, lets the other person book, creates the invite, logs the meeting. Bought because speed to meeting is a conversion lever: the odds of connecting with an inbound enquiry fall sharply with every hour of delay, so routing it straight into a rep's diary is worth money. Chili Piper's specific pitch is that routing — right lead, right rep, instantly. The mild art is knowing when not to send a link: to a senior executive, "would Tuesday 10am or Thursday 2pm suit?" is politer than delegating the work to them.

E-signature — DocuSign, Adobe Acrobat Sign. Sends a document for binding electronic signature, routes multiple signers in order, produces a tamper-evident audit trail. Bought for elapsed time: print-sign-scan-post adds a fortnight to a deal that is otherwise done, and every day between agreement and signature is a day something can go wrong. Is it solid in the UK? For ordinary commercial contracts, yes: electronic signatures are recognised under the Electronic Communications Act 2000 and the UK's retained eIDAS regime, and a 2019 Law Commission report confirmed they can execute documents under English law, including deeds where the formalities are met. Some instruments still have special requirements — that is a question for legal, and knowing it is a question for legal is the correct rep-level knowledge. Daily rep use: prepare the envelope, set signing order, chase politely, and — the detail that separates the organised — confirm in advance who is actually authorised to sign.

Proposals and CPQ. Proposal tools (PandaDoc, Qwilr, Proposify, or native CRM quoting) turn a deal into a branded, trackable document and tell you the prospect spent nine minutes on the pricing page and forwarded it to two colleagues — genuine buying-signal data. CPQ — Configure, Price, Quote — is the heavier machinery: configure a valid combination of products, modules, tiers and terms with rules preventing impossible ones; price it against list, volume tiers, multi-year uplifts and currency rules; quote it and route non-standard discounts through an approval workflow before anything reaches the customer. Bought for margin protection and error prevention: guardrails set once by finance mean every quote is on policy. It sits inside quote-to-cash — quote, contract, order, invoice, revenue recognition — which is why finance cares about it more than you do. The category is in motion (Salesforce moved its long-standing CPQ product to end-of-sale in March 2025, directing new development at Revenue Cloud while existing customers continue on support), but no AE interview will test that. What may be tested is "have you worked with a discount approval process?" — and the right answer explains what an approval matrix is for. Daily rep use: build the quote, anticipate the approval it triggers (a 22% discount needing VP sign-off is a two-day delay to plan for, not discover), and better, shape the deal so no exception is needed.

Revenue intelligence — Clari, Gong Forecast, BoostUp — sits above the CRM and asks whether the forecast is real, comparing rep claims against observed activity: meetings, stakeholder engagement, whether the deal is moving. The practical consequence for a rep is worth internalising: the gap between what you say and what the system observed is now visible.

How it connects. Every tool here either writes back to the CRM or is failing. The connective tissue has names: activity capture (auto-logging emails and meetings from Outlook or Gmail against the right records — the single biggest reduction in CRM drudgery available), bi-directional sync, enrichment, and native integrations versus middleware. The failure mode is a tool that does not sync, creating a second version of the truth and degrading the whole stack to its least reliable component. When a company says "we're consolidating our stack", that is usually what they mean.

One practical note: job adverts list stacks aspirationally. An advert naming seven tools may describe a team whose AEs genuinely live in three. "Which of these do your AEs use daily, and which sit with SDRs or ops?" tells you what your Tuesday looks like and signals that you know a stack list is a shopping list, not a workflow.