A CRM inside Outlook or Gmail

Why the inbox keeps winning

Sales and service people spend most of the working day in mail and calendar. A CRM in a separate tab asks them to duplicate context they already have, and the duplication is where CRMs die. Not because people are lazy, but because typing a summary of an email into another system is obviously redundant work, and obviously redundant work does not survive a busy week.

So the requirement is not really a CRM in Outlook. It is: I should never type the same thing twice, and I should see who this person is without leaving what I am doing. Those are two separate problems and they have different solutions.

What an inbox sidebar actually gives you

The lightweight category, meaning tools that live inside Gmail or Outlook as a panel beside the message. Honest assessment, because for a lot of small teams this is the right purchase.

The pattern is clear enough. Sidebars are excellent at making email into a record and poor at everything that happens after the email. For a two to six person team whose work genuinely is conversations and follow up, that can be the whole job, and paying for anything larger is waste.

Where the inbox model breaks

Four structural limits, none of which better software inside the inbox can solve, because they follow from the mailbox itself.

The mailbox is personal. Its contents belong to one person, and when that person leaves, the history goes with them or becomes an archive nobody searches. A shared record does not have this property, which is exactly why it exists.

Permissions are mailbox permissions. There is no way to let a junior see the client and not the margin, because the sidebar renders whatever the mailbox can reach. Any business with sensitive fields hits this immediately.

Not all work is email. A site visit, a phone call, a signed form, a delivery, a payment. In job and matter shaped businesses most of the real events never touch a mailbox, so an inbox centred system is blind to the majority of what happened.

Reporting needs structure that email does not have. Revenue by source by month requires fields with types. An email thread is prose. You can tag prose, but you cannot sum it.

Put differently: the inbox is a brilliant place to see a record and a poor place to be one. That distinction is the whole design.

The architecture that gives you both

The version we build, and the one every serious CRM converges on: the record lives in the CRM, and the inbox is one of several windows onto it.

Three connections do the work, and they are the same three whichever CRM you end up with.

Email sync, two way. Messages to and from a known address attach themselves to that record automatically, and a reply sent from the CRM arrives from the person's real address and threads correctly in their sent items. Nobody forwards anything to a logging address, because that is a habit and habits fail.

Calendar sync, two way. A meeting booked in either place appears in both, carries the record it relates to, and writes back an outcome. This is the connection that removes double entry for anyone whose day is appointments.

A sidebar in the mail client. Not a second CRM, a window: who this is, their last five interactions, the open work, and one button to log a note or move a stage. The record stays where records belong and the salesperson never leaves the inbox.

Outlook and Microsoft 365 specifics

Worth knowing before anyone promises you anything.

The modern way to connect is Microsoft Graph, which covers mail, calendar, contacts and files across Outlook on the web, Windows, Mac and mobile with one integration. Anything built on it works everywhere the user signs in, which matters more than it sounds.

Two practical points. Permissions are granted by an administrator at the organisation level, so schedule that conversation early rather than discovering it on go live day. And a mailbox integration should be scoped to the folders and the people it needs rather than granted blanket access to every mailbox, which is the default many vendors request and few customers question.

The older desktop add in model, meaning a panel that only exists inside the installed Windows client, is worth avoiding now. It breaks on Mac, does nothing on mobile, and updates at the mercy of the desktop application. If a vendor's Outlook integration is an installed add in only, treat that as a signal about the age of the product.

Gmail and Google Workspace specifics

Similar shape, different mechanics.

Google Workspace add ons run in the sidebar across Gmail on the web and on mobile, using the same code, which is a genuine advantage over the old desktop model. Email and calendar access come through the Gmail and Calendar APIs, again with an administrator granting scopes for the domain.

Two things to check specifically. Scopes that read the full content of mailboxes get extra scrutiny from Google, and any vendor using them should be able to tell you exactly which scopes they request and why. And if your business relies on Google Groups as shared inboxes, confirm how the integration treats them, because messages arriving through a group often do not carry the individual recipient the integration expects, and enquiries quietly fail to attach.

For a Workspace business that only needs contact history beside the message, the sidebar category is strong and cheap. It is when quotes, jobs, scheduling or field staff enter the picture that the model runs out, which is the same boundary as Outlook.

Choosing between the two routes

Our advice is usually sequential rather than either or. Start with a sidebar if your work is genuinely conversations. Move when non email events start mattering, when a second team needs different permissions, or when you cannot answer a question about last quarter without exporting. The full comparison method is in how to choose CRM software.

What we build, and how it feels day to day

In every system we build, email and calendar sync are part of the core work rather than an upgrade, because a CRM that does not know about email is asking people to do the work twice.

In practice that means: a message from a known contact attaches to the right record within seconds, with no forwarding address and no button. A reply sent from the record arrives from the person's own address and appears in their sent items. A meeting booked in Outlook or Google shows on the record with its outcome written back afterwards. And a sidebar in the mail client shows who this is, what is open and what is due, with one click to log a note.

The unglamorous detail that decides whether people trust it: an enquiry from an unknown address does not get silently dropped. It lands in a queue with a suggested match, and someone confirms or creates. Systems that guess quietly are how two records for the same customer appear, and once people stop trusting the matching they go back to their inbox and the CRM becomes decoration.

If you want to see it working before deciding anything, every industry page carries a demo you can click through. Tell us how your team actually works and the quote comes back in writing within two business days, including the case where a sidebar and a spreadsheet are what we would recommend.

The shared inbox problem

Almost every business we work with has at least one shared address: info, sales, bookings, accounts. It is usually the single largest source of enquiries and the single worst managed part of the operation, and no amount of inbox software fixes it, because the problem is not the mail client.

Three failures repeat. Nobody owns a message, so two people reply or nobody does. There is no record of what was promised, because the reply lives in one person's sent items rather than in the thread everyone can see. And the enquiry cannot be attributed to anything, so the marketing that produced it is invisible.

Shared mailbox features help with the first and do nothing for the other two. What actually solves it is treating the shared address as an input into the CRM rather than as a destination. Every message arriving creates or matches a record, gets an owner by rule within seconds, and carries a due task. The reply is sent from the shared address and stored on the record, so the next person who touches that customer sees what was said.

The reporting that follows is the part businesses find most useful, and it is impossible without the record. First response time, distribution of enquiries by hour, and the proportion that got no reply at all, which in the first month of measuring is almost always higher than anybody expected. That number tends to change behaviour faster than any training session.

Tracking, privacy and what to leave switched off

Inbox tools arrive with open tracking, click tracking and sometimes read receipts on by default. They are useful and they carry obligations, so decide deliberately instead of accepting a default.

Open tracking works by embedding a tiny image that reports back when the message is displayed. In the UK and the EU that is generally treated as accessing information on a person's device, which brings it inside the same rules as cookies rather than sitting comfortably inside legitimate interests. Several mail clients now preload images anyway, so a large share of recorded opens are machines rather than people, which makes the data both legally awkward and practically unreliable.

Click tracking is more defensible and more useful, because a click is a deliberate act and the data is accurate. Our default recommendation is straightforward: leave click tracking on, turn open tracking off for one to one sales email, and never let an open drive an automation that a salesperson can see, because it produces the worst possible behaviour, which is calling someone thirty seconds after they opened a message.

Two more settings worth checking on day one. Whether the tool stores full message bodies on its own servers or only references them, because that decides where your customer correspondence physically lives. And what happens to that data when you cancel, since several tools retain synced mail for a period after an account closes. Both answers should be in writing before an administrator grants any scopes.

Logging email automatically, the rule that makes tracking useful

Email tracking in a CRM is only as good as the logging underneath it. If half the team's messages never reach the record, every report about response times and follow ups is wrong, however precise the click data looks.

The rule we set on every build is that logging happens without a decision. Mail to and from any address that matches a contact is attached to that contact automatically, internal mail is ignored, and personal mail is excluded by a short list of domains the user controls. Nobody clicks a log button, because the people who most need logging are the ones who forget to click it.

Once logging is complete, the useful measures are plain ones: time to first reply on a new enquiry, days since last contact on open deals, and quotes with no reply after a week. These come from logged messages and replies, not from opens, which is why they hold up.

A worked example: a moving company quoting by email

A removals firm shows the whole pattern in one workflow, because nearly all of its sales happen by email from an inbox.

1. An enquiry arrives through the website. The CRM creates the lead, sends a survey booking link and assigns it to the estimator covering that postcode.
2. The survey is booked into the estimator's Outlook or Google calendar from the CRM, so the appointment and the customer are linked.
3. The quote goes out from the estimator's own address, built from the survey inventory, and is logged on the record.
4. If there is no reply after three days, the CRM sends a short follow up in the estimator's name. A reply from the customer stops the sequence automatically.
5. On acceptance, the move date goes onto the crew schedule and the customer receives a confirmation with what to prepare.

The customer only ever sees personal email from one person. The business sees every quote, every reply and every lost job in one place. That is the inbox and the CRM working as one, which is the point of this whole guide. The automation page lists the sequences we build most often.