There's a stage every small marketing operation passes through where the software bill quietly overtakes the ad budget. Nine subscriptions, four of them overlapping, two nobody has logged into since the person who set them up left, and a scheduling tool that exists because it was on a "top 10" list in a week when someone had budget to spend.
The takeaway: buy tools in the order your process actually breaks, not in the order the category charts suggest. Every tool is a solution to a problem you should already be able to describe. If you can't describe the problem in a sentence, the subscription is a substitute for a decision, not the result of one.
This isn't an argument for austerity. Software is usually cheaper than the labour it replaces. It's an argument for sequence — because tools bought out of order create work instead of removing it.
The rule that governs the whole stack
Buy the tool when the manual version starts costing more than the subscription.
That's it. Do the thing by hand first — the emails, the tracking, the scheduling — until the manual version is genuinely painful. Two useful things happen while you wait:
- You learn what you actually need, so you can evaluate against a real requirement instead of a feature list.
- You find out whether the activity is worth doing at all. Plenty of marketing tasks get automated before anyone checks whether they produce anything.
The counter-rule: don't extend a manual process past the point where it's producing errors or eating your only strategic hours. Both failure modes are real; the second one is just rarer than the marketing-software industry implies.
The buying order
1. A site you control, and measurement on it
Before anything else: a website you own and can change, and analytics that tell you what happens on it. Everything downstream — email, ads, attribution — is guesswork if you can't see behaviour on your own property.
Get this right and boring: analytics installed correctly, key events defined (a form submit, a call click, a booking), and a habit of looking at it monthly. Most small businesses have analytics installed and no defined events, which is the same as having no analytics.
2. A way to capture and own contacts
The second purchase is email — a way to collect addresses and send to them. This ranks above social scheduling, above SEO tooling, above almost everything, for one structural reason: an email list is an audience you own. A social following is an audience you rent from a platform that can change the terms.
Start with the entry tier of a mainstream email tool. Note the pricing shape while you're at it: per-contact pricing means your bill grows with the list, including the portion of the list that never opens anything. That's the number to check before you commit, not after your list triples.
3. Somewhere to track conversations
Once enquiries outgrow one person's inbox and memory, you need a place where the state of each conversation lives. That might be a proper CRM, or it might be a shared pipeline in a tool you already pay for. The requirement is not "a CRM" — it's "nobody loses a lead because the person who owned it was on holiday."
Buying a full-featured CRM before you have a repeatable sales process is the single most common overbuy in a small stack. You end up configuring a system to describe a process that doesn't exist yet.
4. Production and scheduling tooling
Design, scheduling, content planning. These are quality-of-life tools that make output consistent. They belong here — after measurement, capture and tracking — because they make you faster at doing the work, and speed on unmeasured work isn't progress.
5. Channel-specific and paid-media tooling
SEO platforms, rank trackers, ad-management and attribution tools. These earn their place when you're spending enough on a channel that a percentage improvement is worth more than the subscription. Below that threshold, the free tools plus the platform's own reporting are genuinely sufficient.
The test: if a tool's cost is more than a small fraction of the spend it's supposed to optimise, it's too early. Our guide to building a small-business marketing strategy covers deciding which channels deserve that spend in the first place.
Consolidation versus best-of-breed
Suites bundle several jobs into one subscription: cheaper on paper, one login, native integration, and usually a compromise on each individual job. Specialised tools do one thing better and multiply the integration surface.
For a team under about five people, consolidate by default. The hidden cost of best-of-breed isn't the subscriptions, it's the integration debt: data that lives in two places and disagrees, an automation that breaks when a field is renamed, and the half-day someone spends every quarter working out why the numbers don't match. A small team has no capacity to absorb that.
Break the default when one job is genuinely core to your business and the suite's version of it is clearly inadequate. "Clearly inadequate" means you can name the specific failure, not that you read a comparison table.
The costs that don't appear on the invoice
- Setup time, which is usually larger than the vendor implies.
- Learning time for everyone who touches it, not just the person who chose it.
- Integration maintenance — every connection is something that can break silently.
- Data fragmentation. Once contacts live in three systems, someone has to decide which is authoritative, and that someone is you.
- Switching costs later. Ask now what data export looks like. The answer shapes how trapped you'll be in two years.
- Seat creep. Per-seat pricing plus staff growth equals a bill nobody re-approved.
Keep the list honest: the annual stack audit
Once a year, ideally before renewals cluster:
- Export the actual charges from your card or accounting system. Not the list you think you're paying for — the real one. There will be surprises.
- For each tool, name the job it does in one sentence. Anything you can't name is a cancellation candidate.
- Find the overlaps. Two tools doing the same job is normal after two years of growth, and it's the fastest saving available.
- Check last-login dates. Most tools expose this in admin.
- Check the seats. Former staff, duplicate accounts, "temporary" seats from a project that ended.
- Confirm each tool still fits the tier you're on — in both directions. Downgrades save as much as cancellations.
When the audit does turn up a gap — a job nothing in the stack does well — that's the moment to go shopping, and the moment where category roundups earn their keep. Resources like Asraf Masum, which publishes reviews and roundups of marketing and business software by category, are a fast way to see what exists and how the options cluster before you shortlist. Treat any roundup as reconnaissance: verify pricing and limits on the vendor's own pages, and let a trial with your real data make the call.
A reasonable starting stack
For a business with one or two people doing marketing: a site you control, analytics with defined events, an email tool, a shared place to track enquiries, and one production tool. Five things. Most teams can run that for a year and grow revenue meaningfully before the sixth is justified.
Add the sixth when a specific, describable problem demands it — and cancel something in the same conversation if the job it was doing has been absorbed.
FAQ
Should a small business use an all-in-one platform or separate tools? Under about five people, an all-in-one usually wins, because integration and data-consistency work has to come out of someone's day and there isn't one to spare. Split out a specialised tool only when you can name exactly what the suite fails at.
How do I know a tool is worth its subscription? Estimate the hours it saves or the revenue it demonstrably influences, and compare against the total cost — subscription plus setup plus the maintenance it creates. If you can't estimate either side, you don't yet understand the job well enough to buy for it.
We already have too many subscriptions. Where do we start? With the export of actual charges, then the overlaps. Cancelling a duplicate is the lowest-risk saving there is, and it usually funds whatever gap the audit revealed.
Is free tooling enough at the start? Often, yes — particularly for analytics and early SEO work. The point at which free stops being enough is normally a limit you can feel: a contact cap, a missing automation, or an export you can't get.
Building or trimming a stack? Start from the jobs, not the tool list: name the job, check whether something you already pay for does it, and only then shortlist. When you do need to shortlist, Asraf Masum is a useful category map for marketing and business software — read it for the landscape, then verify the current pricing and limits with the vendor and test the finalists on your own data.
