South African marketing is not short of budget, ambition or technology. It is short of completion.
Every year I go into the research expecting it to tell me that South African marketing teams need more money. More budget, better tools, a clearer mandate from the board. This year the research told me something less comfortable, and it took me a while to stop arguing with it.
We finish far less than we start.
Insight 2027 put 161 questions to 336 respondents, 94.6% of them South African. Across that sample, organisations reported owning 2,848 separate pieces of marketing and sales technology. Of those instances, 1,143 were either not in use at all or in use and not producing a result. That is 40.1%. Two in five of everything the market has bought, signed off, migrated onto and is still paying for is sitting in the corner of the office doing nothing.
I want to be precise about what that figure is and is not. It is not a licence audit. It is not a story about waste, although the waste is real and it is monthly. It is a measure of the distance between a decision and a working capability, and in South African marketing that distance has become the single most expensive thing on the balance sheet.
The money arrived
The easy explanation would be that budgets were cut and teams could not finish what they had started. The data does not support it.
Half of the sample, 50.6%, grew their marketing budget. Once you add the organisations that held flat, 86% either held or grew. Only 14% cut. That is a market with money in it. Whatever is stopping completion, a finance director saying no is not the main culprit.
Nor is appetite. The market is buying technology at a rate that produces 2,848 instances across 336 organisations. It is exploring artificial intelligence enthusiastically enough that on eight of the nine AI functions we tested, a majority of organisations are either not using it or exploring it. Exploration is not the constraint. Money is not the constraint. What is missing sits between the purchase order and the point at which something starts returning.
Where the incompletion shows up
The clearest example in the whole study is the CRM, because it is the one system that everybody agrees matters.
Nearly six in ten in the sample own one, 58.3%. Only 37.2% have one that is working. That leaves 71 organisations that bought a CRM, installed it, and now have something that is not delivering. A dead CRM.
Now look at what happens to their budgets. Among organisations with a working CRM, 61.6% grew their marketing budget. Among organisations with no CRM at all, 45.0% grew. Among organisations with a dead CRM, 42.3% grew. The group that bought the thing and did not land it performs worse than the group that never bought it, and it is not a rounding artefact: the same group also cuts budget more often than anybody else, 19.7% against 12.8% for the working group and 12.1% for those with no CRM.
I have written that finding up on its own in A dead CRM is worse than no CRM, because it deserves the space and because most people’s first reaction is to look for a confound. The obvious one is size, so we tested it. The working-CRM effect holds when you account for company size and for how much of the budget is digital, and size on its own explains none of it. Owning a CRM, on its own, does not predict growth at all. Only having one that works does.
That is the shape of the completion problem in one system. Ownership is not capability. The half-finished state is not a lesser version of the finished state. It is its own condition, and it costs money.
Measurement is where it compounds
If the technology layer is 40.1% idle, the measurement layer is thinner still.
We asked about 14 measures. The average organisation tracks 5.24 of them. Nearly six in ten, 59.8%, say they track return on investment. One in four, 25.0%, can attribute revenue to digital. Both of those cannot be true in the same organisation, and the gap between them is the reason so much technology can sit unused without anybody raising it as a problem. If nobody is measuring the outcome the tool was bought to move, nothing forces the question of whether it landed.
The composite index makes the same point at a structural level. The market scores 49.85 on the infrastructure pillar and 29.27 on measurement. We have bought more than we can see.
This is not a story about small companies
The temptation, if you run a large business, is to read all of this as a small-business problem, and if you run a small business, to read it as the inevitable consequence of not being a large one. Neither survives contact with the data.
The working share of the technology stack runs from 48.8% at start-ups and 50.6% at micro businesses to 67.6% at small businesses, 61.0% at medium, 60.6% at large and 73.5% at enterprise. Landing rates vary enormously, but they do not climb neatly with headcount, and small businesses land a higher share of their stack than large ones do. On the composite capability index, no size band is statistically distinguishable from the market average. Size accounts for three of the 93 associations that survive false-discovery control. Data strategy stage accounts for 31, and digital budget intensity for 30.
So 40.1% is an average, not a law, and I have made that case at length in Forty per cent is an average, not a law. Some organisations get most of their stack working. They are not the biggest ones. They are the ones that finish.
What completion buys you
The part of this I find most useful in client conversations is that the returns to finishing are not linear with effort. They are step changes, and they cluster.
Organisations in the top infrastructure band have 5.13 of nine AI functions in production. Those in the bottom band have 1.26. Nearly three in ten organisations, 29.2%, have no AI function in production at all. Data infrastructure predicts what an organisation has deployed. It tells you nothing about what it is exploring. Everybody explores. Only the organisations with something underneath them ship.
The same pattern appears on channels. Across bands of one to two, three to four, five to six and seven or more channels, working tools rise from 2.91 to 7.98 and measures tracked rise from 3.40 to 7.57. There is no segment in South African marketing that sprays and prays. Breadth arrives with capability, not instead of it.
And the market is not complacent about its own skills. Only 9.2% say they do not keep up with developments in the industry, and that group scores 22.6 on the composite index against 46.0 for everybody else. The appetite to get better is close to universal.
The instruction I would give
If you own the growth number, the useful response to 40.1% is not a procurement review. It is to pick the single system closest to revenue, most likely the CRM, and finish it before you buy anything else. Finish means a definition of a customer that sales and marketing both accept, an owner for each stage, and at least one revenue measure coming out of the other end. That is the argument I make in Finish something before you buy anything, and it is the least glamorous recommendation in the whole programme.
It is also the one with a price attached. The gap between a working CRM and a dead one, on budget growth, is 61.6% against 42.3%. Nothing else in this research moves a commercial outcome that far.
If you want to know more, you can download the full Insight 2027 report.




