Sherlock review

Sherlock is a hosted SaaS product analytics application for scoring engagement, not a library, CLI, or self-hosted server. It aggregates user actions and in-app events, then calculates 1–100 scores for individual users and accounts. Teams choose tracked events and set each event’s weight from 0 to 10; Sherlock applies Winsorizing normalization and exponential distribution algorithms to activity points.

In a SaaS stack, Sherlock sits downstream of product-event collection and upstream of go-to-market workflows. Teams use its rankings, trends, activation tracking, cohorts, and event impact analysis to spot engaged or declining users, compare groups, and assess product usage. Scores and activity can then move into operational tools such as Salesforce, Intercom, Slack, HubSpot CRM, Appcues, and Pipedrive; Copper CRM has a native integration. The product is designed for B2B SaaS teams, including product-led and hybrid sales organizations. It is delivered through a web browser as a cloud service.

The integration surfaces named for Sherlock are Segment for data ingestion and native connections for sending or synchronizing computed scores and activity with Salesforce, Intercom, Slack, HubSpot CRM, Appcues, Pipedrive, and Copper CRM. The vendor lists approximately seven native integrations. This is an integration-led product rather than a documented developer library in the published product details.

Sherlock’s homepage description does not name programming languages, SDKs, a CLI, API protocol, or webhook support. Before building a custom pipeline, confirm in the vendor’s docs whether an API or SDK is available, what event schema it accepts, and how credentials and retries work. The listed integration set can cover common CRM and customer-success workflows, but teams with bespoke event pipelines should verify the ingestion route and any synchronization constraints before committing. Sherlock is a browser-accessible SaaS application, so the described product experience is a hosted dashboard rather than an installable analytics service.

The pricing page lists Team at $49 per month for up to 5 users, unlimited hosts, and a $49 monthly usage credit. Business is $349 per month for up to 100 users, multiple organizations, unlimited hosts, and a $349 monthly usage credit. Enterprise is listed at $4,000 per month, with unlimited users and usage and terms quoted per customer. No free plan or free trial is listed, and annual-billing prices are not published.

The listed overages make this more than a seat-only bill. Beyond included credits, logs and traces cost $0.30 per GB ingested, uncompressed, and $0.10 per GB stored per month; metric samples cost $0.12 per million. Enterprise overage rates are not published. The pricing details list multiple usage-based overage rates but do not give a usage forecast or explain how particular event volumes consume credits. Ask Sherlock to map expected ingestion and retention to a monthly estimate before production, and confirm what happens when a credit is exhausted.

Sherlock is proprietary, closed-source software delivered as a cloud service. There is no open-source edition or self-hosted deployment option, so teams cannot run the scoring system inside their own infrastructure or inspect and modify its code. The unrelated open-source Sherlock Project is a different tool.

The product details describe exporting or synchronizing scores and activity to operational integrations, but do not establish a full data export format or migration path. Before adopting Sherlock as a system of record for engagement, ask how to export raw events, calculated scores, scoring profiles, and historical trends, and whether exports are available without an active subscription. Moving away may also mean rebuilding scoring logic and dashboards elsewhere: event weights and score calculations are part of Sherlock’s hosted product, not a portable library. Treat portability as a procurement question, not an assumed feature.

Sherlock’s listed plans include monthly usage credits and usage-based overages, so production costs depend on ingestion, storage, and metric samples as well as seats. The pricing details do not give a rate-limit schedule, event-volume capacity, retention period, or credit-consumption examples. Ask the vendor for those figures against your expected event load, plus the behavior when ingestion is delayed or a usage credit runs out.

Before depending on scores for account health or sales outreach, confirm ingestion retry semantics, score refresh timing, and how corrections to event data affect historical scores. The product is cloud-only, and the published details do not establish an uptime commitment, security certifications, data residency, or incident-response terms. Get those requirements in writing if Sherlock will process sensitive user or account activity. Also confirm export formats and access to historical data so a service change does not strand the measurements your workflows rely on.