← Help centre

Analytics

Signed in? Ask Lovelio this question inside the app - it answers from this same page.

Analytics is the numbers page for the agency's hiring activity. It lives under "Analytics" in the left navigation and shows seven cards: Floats, Pipeline Funnel, Time-to-Hire, Hiring Velocity, Source Effectiveness, Active Jobs Summary, and Offer Metrics. On top of the page, the AI assistant can answer free-text analytics questions ("how many offers did we send last quarter?") directly in chat, without opening the page.

The analytics page

  • Date range buttons at the top: "Last 7 days", "Last 30 days", "Last 90 days", "All time". The default is last 30 days.
  • Filter dropdowns for Job, Team, and Hiring manager, with a "Clear filters" link when any filter is active. Filters apply to the charts, not to the Active Jobs Summary.
  • An "Export" button with two options: "Save as PDF" (a snapshot image of the whole chart grid, saved as lovelio-analytics.pdf in A4 landscape) and "Export to CSV" (downloads the conversion funnel numbers for the current date range and filters).
  • Access is role-gated: admins, team leaders, and hiring managers see Analytics; roles without analytics access do not see it in the navigation.
  • Empty charts explain what will fill them (for example "No applications in this period") rather than showing a blank panel.

What each card shows

  • Floats: the first card, full width. A belt across the top reads Sent, Opened, Interested, Became a job, Placed for floats sent in the date range, with the previous period of the same length beside it (how many were sent before, and the earlier conversion rate). Below it, two tables: By consultant (sent, opened, interested, jobs, waiting, and the share that became a job) and By team. A float counts in the period it was sent, under the consultant who sent it and the team they were on that day. It counts as a job once it is converted into a job in Lovelio, whenever that happens, so a float sent in June that converted in August is a June win. Waiting means sent or opened with no reply from the client yet, so a low rate on the newest floats reads as unfinished rather than bad. Placed means the converted application became a placement, and the banked figure is the sum of those placements' fees, the same figure as the foot of the Floats board. The Team filter applies; the Job and Hiring manager filters do not (a float has neither) and the card says so when one is set. Who you see follows your role, not the company's visibility setting: an admin sees everyone, a team leader sees their own team, a consultant sees their own line only. The card says which at the top right.
  • Pipeline Funnel: how many applications entered each funnel stage in the period (Maybe, Yes, Submitted, Client Interview, Offer), with the conversion rate from each stage to the next. The Maybe count is the number of applications created in the period.
  • Time-to-Hire: the median number of days from an application being created to it reaching Placed, computed from the real stage-change history of placed candidates. The date range applies to when the application was created, not when it was placed. Below the headline number, a bar per stage shows the median days candidates spent in that stage on their way to being placed. Bars are green under 5 days, amber 5 to 14 days, red over 14 days. Only placed candidates feed this card, so it stays empty until the first placement.
  • Hiring Velocity: a weekly chart (up to 12 weeks) of new applications, stage moves (every stage change counts as a move), and hires. Below it: apps this week, apps this month (last 4 weeks), moves this week and month, hires this month, and a hires trend badge comparing the last 4 weeks to the 4 weeks before.
  • Source Effectiveness: the top 8 sources by application count, showing applications and hires per source, with the hire rate in the hover tooltip. The source is the application's source field; applications with no source show as Unknown.
  • Active Jobs Summary: open jobs, open jobs with no applications, jobs with applications but no interviews scheduled (both flagged amber as warnings), and average applications per job. This card always covers all active jobs and ignores the date range and filters.
  • Offer Metrics: total offers, accepted, declined, pending, acceptance rate, and median days to decision, plus a donut chart of outcomes. There is no offer letter record in Lovelio; an offer means an application moved into the Offer stage. Accepted means it then moved to Placed. Declined means it moved to Rejected or Withdrew after the offer. Pending means it is still sitting at Offer. Acceptance rate is accepted divided by decided offers (accepted plus declined), so pending offers do not drag it down. The date range applies to the offer stage move itself; if the outcome landed after the range ends, the offer shows as pending for that range.

Asking analytics questions in free text

The AI assistant answers any data question over the agency's records with real computed numbers, shown as a table in the chat. Every answer starts with a "Computed:" line saying exactly what was measured, so the numbers are always auditable. The AI never calculates anything itself - a deterministic engine computes every figure from the database.

What can be asked:

  • Money: billings (placement fees, expected GP for contract placements, counted on the candidate's start date, fell-off placements excluded), pending billings, fees at risk in guarantee, average and median fee, median fee rate. Grouped by consultant, each person's money follows the credit split - a 50/50 placement counts half the fee for each.
  • Placements: placements made, fall-offs, fall-off rate.
  • Pipeline: candidates added to jobs, candidates submitted to clients, client and internal interviews, offers, rejections, submission-to-interview rate, interview-to-placement rate, submission-to-placement rate, time to place.
  • Jobs: open jobs, jobs taken on, filled, closed without a fill, fill rate, time to fill. Evergreen pipeline jobs stay out of all of these.
  • Clients: client count, new clients, repeat client rate, days since last touch per client.
  • Activity: candidates added, notes logged, calls logged, emails sent.
  • Floats: floats sent, floats that became jobs, float conversion rate, grouped by client, consultant or team. Same rule as the card: a float counts on the date it was sent and as a job whenever it converted. These follow the normal visibility scope for free-text questions.

How to ask:

  • Group by client, consultant, client industry, job location, work type, source, or stage: "placements by consultant", "billings by industry".
  • Rank and cap: "top ten clients by billing", "bottom 5 clients by submission-to-interview ratio".
  • Time windows: this month, last month, this quarter, last quarter, this year, last year, the last 90 days, or explicit dates. If no window is named, the answer covers all time and the receipt says so.
  • Trends: "billings by month this year".
  • Comparisons: "fill rate this year vs last year".
  • Two-step questions: "of my top 10 clients by billing, which has the worst interview ratio?"

Honesty rules built in:

  • Money is never added up across currencies - totals show one line per currency.
  • Records missing the data a metric needs (a placement with no fee, a filled job with no fill date) are excluded and the answer says how many, in a footnote.
  • The agency's own internal client record is excluded from every figure.
  • If a question is ambiguous, the assistant asks one short question back. If the data does not exist in Lovelio, it says so plainly and the question is logged so the team can close the gap. It never guesses.

Limits and edge cases

  • Time-to-hire and stage medians only exist once at least one candidate has been placed; before that the card shows why it is empty.
  • Stage medians only count candidates who ended up placed, so they describe the winning path, not every candidate.
  • The Hires trend badge needs hires in the previous 4-week period to show a percentage; otherwise it shows a dash.
  • The CSV export covers the conversion funnel; the other cards export via the PDF snapshot.
  • Deleted applications and deleted jobs are excluded from application counts, the funnel, time to hire, sources, and the jobs summary. One exception: the stage-move counts in Hiring Velocity come from the activity history, so moves made on an application that was later deleted can still appear there when no job filter is set.

Common questions

  • Why does my funnel show zero? No applications were created in the selected date range. Widen the range or clear filters.
  • Where do offers come from if there are no offer letters? From stage moves. Moving an application into the Offer stage is what counts as sending an offer.
  • Why does Active Jobs Summary not change when I change the date range? It is a live snapshot of all currently active jobs and deliberately ignores the range and filters.
  • What counts as a hire? An application reaching the Placed stage.
  • Why does the Floats card show a low rate for this week? Floats sent this week have barely had time to land. The Waiting column shows how many are still with the client. Widen the range to judge a rate.
  • Why can I only see my own line on the Floats card? By design. A consultant sees their own floats, a team leader their team, an admin everyone.
  • Can I ask for numbers without opening the page? Yes. Ask the AI assistant, for example "how many placements did each consultant make this quarter" or "show me my top ten clients by billing".
  • Why does a free-text answer differ from a page card? The free-text engine states its exact definition in the "Computed:" line of every answer (for example, billings count on the start date and exclude fell-off placements). A page card may use a different window or definition; the receipt is the place to check.